Multi-parameter comprehensive calibration method of emulsified asphalt thin slurry mixture wet wheel abrasion instrument

Abstract

The invention discloses a multi-parameter comprehensive calibration method for an emulsified asphalt thin slurry mixture wet wheel abrasion instrument, and relates to the technical field of instrument calibration, and the method comprises the following steps: acquiring fault case data of the wet wheel abrasion instrument, analyzing and acquiring key parameter categories corresponding to the calibration of the wet wheel abrasion instrument, and acquiring multi-source key parameters; performing preprocessing on the multi-source key parameters by using variational mode decomposition improved by a grey wolf optimization algorithm; a Bi-LSTM network is adopted to enhance an auto-encoder which introduces noise reduction and sparse functions to construct a comprehensive calibration model, reconstruction data of multi-source key parameters are acquired, and a fault type is determined and an abnormal time period is marked according to reconstruction error distribution; and performing parameter cooperative calibration according to the fault type and the abnormal time period, recording a calibration process, and verifying a calibration effect. According to the invention, the problems of strong parameter coupling, complex dynamic characteristics and the like in the calibration of the wet wheel abrasion tester are solved, and the accuracy of the abrasion resistance test of the emulsified asphalt thin slurry mixture is obviously improved.

Description

Technical Field

[0001] The present invention relates to the technical field of instrument calibration, and more specifically to a multi-parameter comprehensive calibration method for an emulsified asphalt slurry mixture wet wheel abrasion tester. Background Art

[0002] Emulsified asphalt slurry seal mixture (Emulsified Asphalt Slurry Seal Mixture) is an economical and environmentally friendly pavement maintenance material widely used for preventive maintenance and minor damage repair. Its performance evaluation primarily includes its abrasion resistance, water damage resistance, and asphalt-aggregate compatibility. The wet track abrasion test (WTAT) is a key method for evaluating the durability of slurry mixtures. The wet track abrasion test simulates wheel load and water erosion, measuring the mass loss of specimens under specific conditions to assess the mixture's abrasion resistance.

[0003] Currently, the calibration of wet-wheel abraders primarily relies on single-parameter verification (such as speed, wear time, and rubber wheel pressure), lacking a systematic, multi-parameter coordinated calibration method. Because wet-wheel abrasion test results are affected by multiple factors, including the instrument's operating status, environmental conditions, and operating specifications, relying solely on single-parameter calibration can lead to test data deviations, affecting the accurate determination of the mixture's optimal asphalt content (OAC), and thus reducing the pavement performance of the slurry seal. Furthermore, wet-wheel abraders produced by different manufacturers differ in structural design and loading methods. Inconsistent calibration methods can lead to incomparable test results, impacting engineering applications. Therefore, establishing a wet-wheel abrader calibration method based on multi-parameter comprehensive calibration, encompassing key parameters such as speed accuracy, rubber wheel wear status, water bath temperature stability, wear time control, and loading pressure uniformity, is crucial for improving the accuracy and repeatability of test data. Summary of the Invention

[0004] In order to solve the above technical problems, the present invention proposes a multi-parameter comprehensive calibration method for the wet wheel abrasion tester of emulsified asphalt slurry mixture, which improves the accuracy of the anti-wear performance test of emulsified asphalt slurry mixture and provides a reliable basis for slurry mixture mix design, quality control and project acceptance.

[0005] A first aspect of the present invention provides a multi-parameter comprehensive calibration method for an emulsified asphalt slurry mixture wet wheel abrasion tester, comprising the following steps: Obtaining fault case data of the wet wheel abrasion tester, preprocessing the fault case data, analyzing and obtaining key parameter categories corresponding to the calibration of the wet wheel abrasion tester, and collecting multi-source key parameters according to the key parameter categories; The variational modal decomposition is improved based on the Grey Wolf Optimization Algorithm, and the multi-source key parameters are preprocessed using the improved variational modal decomposition; A comprehensive calibration model is constructed using an autoencoder enhanced with a Bi-LSTM network and introduced with noise reduction and sparsity functions. The preprocessed multi-source key parameters are imported into the comprehensive calibration model to obtain reconstructed data of the multi-source key parameters. The fault type is determined based on the reconstruction error distribution and the abnormal period is marked. Perform parameter collaborative calibration according to the fault type and abnormal period, record the calibration process, and verify the calibration effect.

[0006] In this solution, the fault case data of the wet wheel abrasion tester is obtained, the fault case data is preprocessed, and the key parameter categories corresponding to the calibration of the wet wheel abrasion tester are analyzed and obtained, specifically: Extracting fault case data based on historical test data, historical operation and maintenance data, and fault simulation data of the emulsified asphalt slurry mixture wet wheel abrasion tester, performing data cleaning on the fault case data, and labeling the pre-processed fault case data using the fault type; Conduct parameter tracing on fault case data, extract the parameters involved in the fault, and classify the extracted parameters into continuous parameters, discrete parameters, and periodic parameters. Perform regional isolation forest modeling on different parameters, and introduce a hypersphere partitioning strategy to define the data space. Calculate the path length based on the number of spheres passed from the root node to the leaf node, calculate the parameter importance based on the path length using the characteristic perturbation method, screen parameters that meet the preset importance threshold, and sort the screened parameters according to the parameter importance; According to the sorting results, a preset number of parameters are selected as key parameter categories corresponding to the wet wheel abrasion tester calibration, and multi-source key parameters are collected based on the key parameter categories.

[0007] In this solution, the variational modal decomposition is improved based on the Grey Wolf Optimization Algorithm, and the multi-source key parameters are preprocessed using the improved variational modal decomposition, specifically: Initialize the parameters of the gray wolf optimization algorithm, set the range of the number of modes and the penalty factor, randomly initialize the position vector of each gray wolf, select the quality factor to construct the fitness function, and perform variational mode decomposition for each parameter combination; The top three best solutions are selected as α wolf, β wolf, and δ wolf according to fitness ranking, and the current global optimal solution is recorded. In the iterative calculation, the positions of α wolf, β wolf, δ wolf and other gray wolves are updated using the steps of surrounding, hunting and attacking prey. The gray wolves before and after the position update are paired with each other, and uniform crossover is performed according to the preset probability to generate a new gray wolf population; The fitness of the gray wolves in the new gray wolf population is calculated, and a mutation operator that changes with the number of iterations is introduced to perform Gaussian mutation on the α wolf with the highest fitness. The fitness of the α wolf before and after the mutation is compared, and the elite retention strategy is used to select the better solution to enter the next generation. When the number of iterations reaches the maximum number of iterations or the iteration termination condition is met, the position of the α wolf with the highest fitness in the last iteration is output, and the optimal mode number and penalty factor corresponding to the multi-source key parameters are obtained; The optimal mode number and penalty factor corresponding to the multi-source key parameters are used to configure the variational mode decomposition for parameter adaptive decomposition, calculate the sample entropy of each IMF component, retain the components that meet the preset entropy value range, and obtain the preprocessed multi-source key parameters.

[0008] In this solution, a Bi-LSTM network is used to enhance the autoencoder with noise reduction and sparsity functions to build a comprehensive calibration model. Specifically: A dual-channel hybrid neural network structure is used to construct a comprehensive calibration model, which includes a timing extraction layer and a feature optimization layer. In the feature extraction layer, a Bi-LSTM network and a noise reduction encoder are used to construct a Bi-LSTM noise reduction encoder; Adaptive noise is added to the input multi-source key parameters according to the corresponding parameter importance. A Bi-LSTM denoising encoder is used to perform bidirectional feature extraction on the noisy multivariate time series. A spatiotemporal attention mechanism is introduced to obtain the weighted fusion of spatiotemporal attention weights. Adversarial training and Gaussian masking strategies are used for denoising. In the feature optimization layer, a sparse autoencoder is used to compress the feature space and retain key information, the features output by the Bi-LSTM denoising encoder are sparsely represented using the KL divergence penalty term, and the independence of features in each dimension is enhanced through orthogonal regularization; An inverse Bi-LSTM network is used to construct the decoder of the comprehensive calibration model. Skip connections are introduced to retain high-frequency details. The importance weights of each time step are calculated to focus on reconstructing fault-sensitive periods. Weighted time series features are used to guide the reconstruction of multi-source key parameters. The reconstructed multi-source key parameters are imported into the output layer, and a multi-task head structure is used to locate the fault type.

[0009] In this solution, the pre-processed multi-source key parameters are imported into the comprehensive calibration model to obtain the reconstructed data of the multi-source key parameters. The fault type is determined based on the reconstruction error distribution and the abnormal period is marked. Specifically: After preprocessing, the key parameters of multiple sources are standardized and time-aligned, and a multi-dimensional time series matrix is ​​constructed as the input of the comprehensive calibration model. A Bi-LSTM denoising encoder is used to extract bidirectional time series features and perform spatiotemporal attention weighting. The features output by the Bi-LSTM denoising encoder are imported into the sparse autoencoder for compression and reconstruction. The reconstructed data is used to obtain the time domain error, frequency domain error and feature space error in the bottleneck layer, generating a multi-dimensional reconstruction error. A dynamic threshold is set based on the historical normal data of the wet wheel abrasion tester of the emulsified asphalt slurry mixture, and an adaptive sliding window is set according to the short-term anomaly detection and the long-term anomaly detection. The adaptive sliding window is used to determine whether the multi-dimensional reconstruction error meets the dynamic threshold, and the abnormal period detection is performed based on the judgment result. A principal component analysis is performed based on the multi-dimensional reconstruction error, and the principal component reconstruction error is obtained as the principal component direction for principal component projection to generate a reconstruction error distribution. The fault feature pattern is matched using the Mahalanobis distance according to the abnormal period detection result and the reconstruction error distribution, and the fault category is identified according to the matching result.

[0010] In this scheme, the threshold value is adjusted dynamically according to the aging of the wet wheel abrasion tester of emulsified asphalt slurry mixture, specifically: Obtain the baseline normal data of the multi-source key parameters of the emulsified asphalt slurry mixture wet wheel abrasion tester, define the aging coefficient vector of the multi-source key parameters, and construct an aging model based on the baseline normal data, aging coefficient vector, and equipment usage intensity and usage time; In the aging model, a particle swarm algorithm is used to optimize the aging coefficient, the value range of the aging coefficient is initialized, and the fitness of the particle is calculated according to the deviation between the predicted normal data and the actual normal data of the current multi-source key parameters; The chaotic algorithm is introduced to perform chaotic processing on the fitness of the current particle, and the optimal position moving direction is determined by using the differential idea to update the position. The optimal solution of the aging coefficient is obtained by iterative position update until the maximum number of iterations is reached. The optimal solution of the aging coefficient is used to weight the threshold corresponding to the historical normal data and dynamically adjust the threshold.

[0011] In this solution, parameter collaborative calibration is performed according to the fault type and abnormal period, and the calibration process is recorded and the calibration effect is verified. Specifically: Preset a hierarchical response mechanism according to different fault types, query and set a calibration priority in the hierarchical response mechanism according to the current fault type and abnormal period, and determine the calibration time based on the calibration priority; Obtain the error profiles corresponding to the main key parameters and associated key parameters of the current fault type, introduce a stacking strategy to build a retrieval model, train a KNN model based on the distance feature vectors corresponding to Euclidean distance, Manhattan distance, and Fréchet distance, output a preset number of calibration instances, and score them based on the calibration results. Use the scoring results to determine the compensation coefficients for the main key parameters and associated key parameters. Use a preset electronic form to record the calibration process, compare the difference values ​​of the main key parameters before and after calibration, and establish a control chart to track the difference values ​​for stability monitoring and verification.

[0012] The second aspect of the present invention provides a multi-parameter comprehensive calibration system for an emulsified asphalt slurry mixture wet wheel abrasion meter, comprising: a memory and a processor, wherein the memory includes a multi-parameter comprehensive calibration method program for an emulsified asphalt slurry mixture wet wheel abrasion meter. When the multi-parameter comprehensive calibration method program for an emulsified asphalt slurry mixture wet wheel abrasion meter is executed by the processor, the steps of the multi-parameter comprehensive calibration method for an emulsified asphalt slurry mixture wet wheel abrasion meter are implemented.

[0013] Compared with the prior art, the present invention has the following beneficial effects: The present invention solves the difficult problems of strong parameter coupling and complex dynamic characteristics in the calibration of wet wheel abrasion testers, and significantly improves the accuracy of the anti-wear performance test of emulsified asphalt slurry mixtures. In the technical solution of the present invention, an improved variational modal decomposition is used to adaptively decompose non-stationary signals such as rotation speed, temperature, and pressure, effectively separating noise and fault characteristics, reducing the calibration error of key parameters, and the LSTM-VAE model accurately captures the time dependence of dynamic parameters, improving calibration accuracy and reliability. Through latent space mapping, complex chain failure modes are automatically identified, multi-parameter collaborative optimization is achieved, and the risk of underreporting is greatly reduced. Adaptive calibration and closed-loop calibration are used to improve engineering applicability and automation level. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] In order to more clearly illustrate the technical solutions in the embodiments or exemplary embodiments of the present invention, the following briefly introduces the drawings required for use in the embodiments or exemplary descriptions. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained according to these drawings without paying any creative work.

[0015] Figure 1 A flow chart showing a multi-parameter comprehensive calibration method for a wet wheel abrasion tester for an emulsified asphalt slurry mixture is provided; Figure 2 A flowchart of constructing a comprehensive calibration model in an embodiment is shown; Figure 3 A flowchart illustrating dynamically adjusting thresholds according to device aging in an embodiment is shown; Figure 4 The block diagram of the multi-parameter comprehensive calibration system of the emulsified asphalt slurry mixture wet wheel abrasion tester is shown. DETAILED DESCRIPTION

[0016] In order to more clearly understand the above-mentioned objects, features and advantages of the present invention, the present invention is further described in detail below in conjunction with the accompanying drawings and specific embodiments. It should be noted that, in the absence of conflict, the embodiments of the present application and the features therein can be combined with each other.

[0017] In the following description, many specific details are set forth to facilitate a full understanding of the present invention. However, the present invention may also be implemented in other ways different from those described herein. Therefore, the scope of protection of the present invention is not limited to the specific embodiments disclosed below.

[0018] Figure 1 A flow chart showing a multi-parameter comprehensive calibration method for a wet wheel abrasion tester for emulsified asphalt slurry mixture is provided.

[0019] like Figure 1 As shown, this embodiment provides a multi-parameter comprehensive calibration method for an emulsified asphalt slurry mixture wet wheel abrasion tester, comprising: S102, obtaining fault case data of the wet wheel abrasion tester, preprocessing the fault case data, analyzing and obtaining key parameter categories corresponding to the calibration of the wet wheel abrasion tester, and collecting multi-source key parameters according to the key parameter categories; S104, improving variational mode decomposition based on the Grey Wolf Optimization Algorithm, and preprocessing the multi-source key parameters using the improved variational mode decomposition; S106, using a Bi-LSTM network to enhance an autoencoder with noise reduction and sparsity functions to build a comprehensive calibration model, importing the preprocessed multi-source key parameters into the comprehensive calibration model, obtaining reconstructed data of the multi-source key parameters, determining the fault type based on the reconstruction error distribution, and marking the abnormal period; S108, performing parameter collaborative calibration according to the fault type and abnormal period, recording the calibration process, and verifying the calibration effect.

[0020] It should be noted that fault case data is extracted based on the historical test data, historical operation and maintenance data and fault simulation data of the emulsified asphalt slurry mixture wet wheel abrasion tester, the fault case data is cleaned, and the pre-processed fault case data is labeled with the fault type, such as mechanical system failure (bearing wear, belt slip), control system failure (PID imbalance, sensor drift), measurement system failure (uneven pressure, timing error), etc.; the fault case data is traced and the parameters involved in the failure are extracted. The extracted parameters are divided into continuous parameters, discrete parameters and periodic parameters, and regional isolation forest modeling is performed on different parameters. The hypersphere partitioning strategy is introduced to define the data space, and local sensitive areas are formed through radius constraints. For continuous parameters such as speed, an adaptive radius hypersphere is used, in which the radius constraint automatically adjusts the size of the hypersphere according to the parameter distribution density; for discrete parameters such as wear state, a fixed radius hypersphere is used; for periodic parameters such as temperature, an annular partitioning with angle constraints is introduced.

[0021] The path length is calculated based on the number of spheres passed from the root node to the leaf node. The characteristic contribution is defined based on the path length using the characteristic perturbation method. The parameter importance is calculated, and parameters that meet a preset importance threshold are screened. The screened parameters are then ranked according to the parameter importance. Preferably, key parameters are screened using multiple criteria, stability is verified using bootstrap sampling, engineering relevance is calculated based on the Spearman correlation coefficient with wear volume, and multi-criteria screening is performed using importance, stability, and engineering relevance. Based on the ranking results, a preset number of parameters are selected as key parameter categories corresponding to the wet wheel abrader calibration, and multi-source key parameters are collected based on the key parameter categories. Preferably, the calibration of the wet wheel abrader involves speed accuracy, rubber wheel wear state, water bath temperature stability, wear time control and loading pressure uniformity; a high-precision photoelectric encoder is used to measure the actual speed, and the speed accuracy is obtained by comparing it with the set value; the surface morphology of the rubber wheel is obtained by a 3D profile scanner, and the wear depth and uniformity are calculated to obtain the rubber wheel wear state; a high-precision temperature sensor is used to record water bath temperature fluctuations, and the temperature control performance is evaluated to obtain water bath temperature stability; a high-resolution timer is used to measure the actual wear time, and the wear time control is obtained by comparing it with the set time; a pressure sensor is arranged on the contact surface between the rubber wheel and the specimen to detect the uniformity of pressure distribution to obtain the uniformity of loading pressure.

[0022] It should be noted that an improved gray wolf optimization algorithm is used to intelligently optimize the two key parameters of variational modal decomposition (the number of modes and the penalty factor). The optimal parameter combination is obtained through multi-stage iterative calculations. The parameters of the gray wolf optimization algorithm are initialized, the range of the mode number and the penalty factor is set, the gray wolf population size is set, the maximum number of iterations is defined, and the position vector of each gray wolf is randomly initialized to ensure that the mode number is an integer and the penalty factor is within the set range. A quality factor is selected to construct a fitness function, and variational modal decomposition is performed for each parameter combination. The quality factor of each IMF component is calculated as (signal energy / noise energy) × (main frequency bandwidth / total bandwidth). The fitness function is constructed based on the weighted sum of the quality factors of all IMF classifications. According to the fitness ranking, the top three optimal solutions are selected as α wolf, β wolf, and δ wolf, and the current global optimal solution is recorded. In the iterative calculation, the steps of surrounding, hunting and attacking prey are used to update the positions of α wolf, β wolf, δ wolf and other gray wolves. The gray wolves before and after the position update are paired with each other, and uniform crossover is performed according to the preset probability. The parameter values ​​of the modal number and the penalty factor are preferably exchanged with a probability of 0.5 to ensure that the newly generated modal number is an integer to generate a new gray wolf population; the fitness of the gray wolf is calculated in the new gray wolf population, and a mutation operator that changes with the number of iterations is introduced to perform Gaussian mutation on the α wolf with the highest fitness. Mutation, with the probability of mutation decreasing as the number of iterations increases, compares the fitness of the α wolf before and after mutation, and uses an elite retention strategy to select the best solution for the next generation. When the number of iterations reaches the maximum number of iterations or the iteration termination condition is met, the position of the α wolf with the highest fitness in the last iteration is output, and the optimal number of modes and penalty factors corresponding to the multi-source key parameters are obtained. The optimal number of modes and penalty factors corresponding to the multi-source key parameters are used to configure variational mode decomposition for parameter adaptive decomposition, calculate the sample entropy of each IMF component, retain the components that meet the preset entropy value range, and obtain the preprocessed multi-source key parameters. By combining the intelligent optimization algorithm with adaptive signal decomposition, a precise feature extraction foundation is provided for the multi-parameter collaborative calibration of the wet wheel wear tester.

[0023] Figure 2 A flowchart of constructing a comprehensive calibration model in an embodiment is shown.

[0024] According to an embodiment of the present invention, a comprehensive calibration model is constructed by using a Bi-LSTM network to enhance the autoencoder with noise reduction and sparsity functions, specifically: S202, constructing a comprehensive calibration model using a dual-channel hybrid neural network structure, wherein the comprehensive calibration model includes a time series extraction layer and a feature optimization layer, and constructing a Bi-LSTM noise reduction encoder using a Bi-LSTM network and a noise reduction encoder in the feature extraction layer; S204: Adaptive noise is added to the input multi-source key parameters according to the importance of the corresponding parameters. A Bi-LSTM noise reduction encoder is used to perform bidirectional feature extraction on the noisy multivariate time series. A spatiotemporal attention mechanism is introduced to obtain a weighted fusion of spatiotemporal attention weights. Adversarial training and Gaussian masking strategies are used for noise reduction. S206, using a sparse autoencoder in the feature optimization layer to compress the feature space and retain key information, sparsely representing the features output by the Bi-LSTM denoising encoder using a KL divergence penalty term, and enhancing the independence of features in each dimension through orthogonal regularization; S208 uses an inverse Bi-LSTM network to build a decoder for the comprehensive calibration model, introduces skip connections to retain high-frequency details, calculates the importance weights of each time step to focus on reconstructing fault-sensitive periods, uses weighted time series features to guide the reconstruction of multi-source key parameters, imports the reconstructed multi-source key parameters into the output layer, and uses a multi-task head structure to locate the fault type.

[0025] It should be noted that a comprehensive calibration model is constructed by using a Bi-LSTM network to enhance the autoencoder with noise reduction and sparsity functions. The bidirectional long short-term memory network captures the temporal dependencies, the noise reduction coding enhances the robustness to sensor noise, and the sparsity constraint avoids overfitting and improves feature interpretability. The training process of the comprehensive calibration model is divided into two stages: first, pre-training the Bi-LSTM denoising encoder, and then end-to-end fine-tuning the entire network. After pre-processing, the multi-source key parameters are standardized and time-aligned. A multidimensional time series matrix is ​​constructed as the input of the comprehensive calibration model. The Bi-LSTM denoising encoder is used to extract bidirectional time series features and perform spatiotemporal attention weighting, focusing on abnormality-sensitive periods. The features output by the Bi-LSTM denoising encoder are imported into a sparse autoencoder for compression and reconstruction. The bottleneck layer extracts features and then uses the decoder to restore the original dimensions. The reconstructed data is used to obtain the time domain error, the frequency domain error after FFT transformation, and the feature space error in the bottleneck layer to generate a multidimensional reconstruction error. A dynamic threshold is set based on the historical normal data of the wet wheel abrasion tester of emulsified asphalt slurry mixture. An adaptive sliding window is set based on short-term and long-term anomaly detection. The adaptive sliding window is used to determine whether the multidimensional reconstruction error meets the dynamic threshold. The abnormal period is detected based on the judgment result. Short-term anomalies (<5 seconds) are detected, such as instantaneous speed fluctuations and pressure drops. Long-term anomalies (>30 seconds) include wear trend deviations and temperature drift. Principal component analysis is performed based on the multi-dimensional reconstruction error. The principal component reconstruction error is used as the principal component direction for principal component projection to generate a reconstruction error distribution. Based on the abnormal period detection results and the reconstruction error distribution, fault feature pattern matching is performed using Mahalanobis distance. The fault type is identified based on the matching results. For example, the characteristic pattern corresponding to belt slip is high-frequency speed burrs, which typically occur during the acceleration phase.

[0026] Figure 3 A flow chart of dynamically adjusting thresholds according to device aging in an embodiment is shown.

[0027] According to an embodiment of the present invention, the threshold value is dynamically adjusted according to the aging of the wet wheel abrasion tester for emulsified asphalt slurry mixture, specifically: S302, obtaining baseline normal data corresponding to multi-source key parameters of the emulsified asphalt slurry mixture wet wheel abrasion tester, defining an aging coefficient vector for the multi-source key parameters, and constructing an aging model based on the baseline normal data, the aging coefficient vector, and the equipment usage intensity and usage time; S304, using a particle swarm algorithm to optimize the aging coefficient in the aging model, initializing the value range of the aging coefficient, and calculating the fitness of the particles based on the deviation between the predicted normal data and the actual normal data of the current multi-source key parameters; S306, introducing a chaotic algorithm to perform chaotic processing on the fitness of the current particle, and using the differential idea to determine the optimal position movement direction for position update, and iterating the position update until the maximum number of iterations is reached to obtain the optimal solution of the aging coefficient; S308: Use the optimal solution of the aging coefficient to weight the threshold corresponding to the historical normal data and dynamically adjust the threshold.

[0028] It should be noted that during the acceptance period of new equipment (the first three months), multi-source key parameters are collected, abnormal operating condition data are eliminated, statistical features are extracted, and baseline normal data are obtained. The aging effect equation in the aging model is expressed as , Indicates the equipment usage intensity coefficient, represents the aging coefficient, Indicates usage time. represents the baseline normal data, Indicates the normal data after aging, and the predicted normal data based on the current multi-source key parameters Compared with actual normal data The fitness of the particle is calculated by the deviation of , and the fitness function is: In the particle swarm algorithm, logistic mapping is used to generate the initial population, and a differential strategy is introduced to improve the convergence speed and global search success rate. Intelligent threshold adjustment is achieved by quantifying the aging status of the equipment. In addition, maintenance decision support is provided based on the quantified aging status of the equipment. When the aging coefficient of a key parameter among the multi-source key parameters is greater than the preset threshold, an operation and maintenance warning is generated. In addition, the aging model is used to predict the next calibration time, and warning information is generated before the next calibration time.

[0029] It should be noted that a hierarchical response mechanism is preset based on different fault types. The preferred hierarchical response mechanism includes immediate calibration (P0): for faults that directly affect test accuracy, such as temperature runaway and abnormal speed; planned calibration (P1): for progressive faults such as wear and uneven pressure; and observation and monitoring (P2): for abnormal auxiliary parameters such as ambient humidity. The calibration priority is set in the hierarchical response mechanism based on the current fault type and abnormal time period, and the calibration time is determined based on the calibration priority. The error profile corresponding to the main key parameters and associated key parameters of the current fault type is obtained. For example, the main key parameters of bearing wear are speed fluctuation and vibration, and the associated key parameter is wear uniformity. The preferred error profile includes time domain features, frequency domain features, and morphological features obtained through a dynamic time warping algorithm. A stacking strategy is introduced to construct a retrieval model. A KNN model is trained based on the distance feature vectors corresponding to the Euclidean distance, Manhattan distance, and Fréchet distance. A preset number of calibration examples are output, and a score is calculated based on the calibration effect. The score results are used to determine the compensation coefficients for the main key parameters and associated key parameters. The calibration process is recorded using a preset electronic form. The preferred preset electronic form includes the calibration time, fault code, pre-compensation deviation, compensation coefficient, etc. The difference values ​​of the main key parameters before and after calibration are compared, and an X-bar-R control chart is established to track the difference values ​​for stability monitoring and verification. The stability criteria include 7 consecutive points on the same side of the center line, any point exceeding the control limit, and obvious non-random patterns. Closed-loop calibration is integrated into the wet wheel abrasion tester hardware to support unattended automatic calibration every 24 hours, reducing the need for manual intervention. In addition, a self-learning calibration knowledge base is constructed to structure the storage of typical fault calibration schemes. Closed-loop feedback is formed during each calibration. When the calibration verification passes, it is stored in the calibration knowledge base. The calibration knowledge base is automatically clustered and updated every month to obtain the multi-dimensional reconstruction error of the multi-source key parameters of the current emulsified asphalt slurry mixture wet wheel abrasion tester, the error portraits corresponding to the main key parameters and the related key parameters. The multi-dimensional reconstruction error, the error portraits corresponding to the main key parameters and the related key parameters are used to perform similarity calculation in the calibration knowledge base to obtain the fault calibration scheme that meets the similarity threshold. The current calibration scheme is generated based on the screened fault calibration schemes. In addition, the remaining life is predicted by the control chart trend, and a predictive maintenance warning is generated based on the remaining life.

[0030] Figure 4 The block diagram of the multi-parameter comprehensive calibration system of the emulsified asphalt slurry mixture wet wheel abrasion tester is shown.

[0031] The second embodiment of the present invention provides a multi-parameter comprehensive calibration system 4 for an emulsified asphalt slurry mixture wet wheel abrasion meter, comprising: a memory 41 and a processor 42. The memory includes a multi-parameter comprehensive calibration method program for an emulsified asphalt slurry mixture wet wheel abrasion meter. When the multi-parameter comprehensive calibration method program for an emulsified asphalt slurry mixture wet wheel abrasion meter is executed by the processor, the steps of the multi-parameter comprehensive calibration method for an emulsified asphalt slurry mixture wet wheel abrasion meter are implemented.

[0032] The third aspect of the present invention provides a computer-readable storage medium, which includes a multi-parameter comprehensive calibration method program for an emulsified asphalt slurry mixture wet wheel abrasion meter. When the multi-parameter comprehensive calibration method program for an emulsified asphalt slurry mixture wet wheel abrasion meter is executed by a processor, the steps of the multi-parameter comprehensive calibration method for an emulsified asphalt slurry mixture wet wheel abrasion meter are implemented.

[0033] In the several embodiments provided in this application, it should be understood that the disclosed methods and systems can be implemented in other ways. The system embodiments described above are merely illustrative. For example, the division of the units is merely a logical function division. In actual implementation, there may be other division methods, such as: multiple units or components can be combined, or can be integrated into another system, or some features can be ignored, or not executed. In addition, the coupling, direct coupling, or communication connection between the components shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms. In addition, the functional units in the various embodiments of the present invention can all be integrated into one processing unit, or each unit can be a separate unit, or two or more units can be integrated into one unit; the above-mentioned integrated units can be implemented in the form of hardware or in the form of hardware plus software functional units.

[0034] Those skilled in the art will understand that all or part of the steps of the above-mentioned method embodiment can be completed by hardware related to program instructions, and the aforementioned program can be stored in a computer-readable storage medium. When the program is executed, it executes the steps of the above-mentioned method embodiment; and the aforementioned storage medium includes: mobile storage devices, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disks or optical disks, and other media that can store program codes.

[0035] The above description is only a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any technician familiar with this technical field can easily think of changes or replacements within the technical scope disclosed by the present invention, which should be covered by the scope of protection of the present invention.

Claims

1. A multi-parameter comprehensive calibration method for a wet wheel abrasion tester for an emulsified asphalt slurry mixture, characterized in that: The following steps are involved: Obtaining fault case data of the wet wheel abrasion tester, preprocessing the fault case data, analyzing and obtaining key parameter categories corresponding to the calibration of the wet wheel abrasion tester, and collecting multi-source key parameters according to the key parameter categories; The variational modal decomposition is improved based on the Grey Wolf Optimization Algorithm, and the multi-source key parameters are preprocessed using the improved variational modal decomposition; A comprehensive calibration model is constructed using an autoencoder enhanced with a Bi-LSTM network and introduced with noise reduction and sparsity functions. The preprocessed multi-source key parameters are imported into the comprehensive calibration model to obtain reconstructed data of the multi-source key parameters. The fault type is determined based on the reconstruction error distribution and the abnormal period is marked. Perform parameter collaborative calibration according to the fault type and abnormal period, record the calibration process, and verify the calibration effect.

2. The multi-parameter comprehensive calibration method for the wet wheel abrasion tester of emulsified asphalt slurry mixture according to claim 1 is characterized in that: Obtain the fault case data of the wet wheel abrasion tester, pre-process the fault case data, and analyze and obtain the key parameter categories corresponding to the calibration of the wet wheel abrasion tester, specifically: Extracting fault case data based on historical test data, historical operation and maintenance data, and fault simulation data of the emulsified asphalt slurry mixture wet wheel abrasion tester, performing data cleaning on the fault case data, and labeling the pre-processed fault case data using the fault type; Conduct parameter tracing on fault case data, extract the parameters involved in the fault, and classify the extracted parameters into continuous parameters, discrete parameters, and periodic parameters. Perform regional isolation forest modeling on different parameters, and introduce a hypersphere partitioning strategy to define the data space. Calculate the path length based on the number of spheres passed from the root node to the leaf node, calculate the parameter importance based on the path length using the characteristic perturbation method, screen parameters that meet the preset importance threshold, and sort the screened parameters according to the parameter importance; According to the sorting results, a preset number of parameters are selected as key parameter categories corresponding to the wet wheel abrasion tester calibration, and multi-source key parameters are collected based on the key parameter categories.

3. The multi-parameter comprehensive calibration method for the wet wheel abrasion tester of emulsified asphalt slurry mixture according to claim 1 is characterized in that: The variational modal decomposition is improved based on the Grey Wolf Optimization Algorithm, and the multi-source key parameters are preprocessed using the improved variational modal decomposition, specifically: Initialize the parameters of the gray wolf optimization algorithm, set the range of the number of modes and the penalty factor, randomly initialize the position vector of each gray wolf, select the quality factor to construct the fitness function, and perform variational mode decomposition for each parameter combination; The top three best solutions are selected as α wolf, β wolf, and δ wolf according to fitness ranking, and the current global optimal solution is recorded. In the iterative calculation, the positions of α wolf, β wolf, δ wolf and other gray wolves are updated using the steps of surrounding, hunting and attacking prey. The gray wolves before and after the position update are paired with each other, and uniform crossover is performed according to the preset probability to generate a new gray wolf population; The fitness of the gray wolves in the new gray wolf population is calculated, and a mutation operator that changes with the number of iterations is introduced to perform Gaussian mutation on the α wolf with the highest fitness. The fitness of the α wolf before and after the mutation is compared, and the elite retention strategy is used to select the better solution to enter the next generation. When the number of iterations reaches the maximum number of iterations or the iteration termination condition is met, the position of the α wolf with the highest fitness in the last iteration is output, and the optimal mode number and penalty factor corresponding to the multi-source key parameters are obtained; The optimal mode number and penalty factor corresponding to the multi-source key parameters are used to configure the variational mode decomposition for parameter adaptive decomposition, calculate the sample entropy of each IMF component, retain the components that meet the preset entropy value range, and obtain the preprocessed multi-source key parameters.

4. The multi-parameter comprehensive calibration method for the wet wheel abrasion tester of emulsified asphalt slurry mixture according to claim 1 is characterized in that: A comprehensive calibration model is constructed by using a Bi-LSTM network to enhance the autoencoder with noise reduction and sparsity functions. Specifically: A dual-channel hybrid neural network structure is used to construct a comprehensive calibration model, which includes a timing extraction layer and a feature optimization layer. In the feature extraction layer, a Bi-LSTM network and a noise reduction encoder are used to construct a Bi-LSTM noise reduction encoder; Adaptive noise is added to the input multi-source key parameters according to the corresponding parameter importance. A Bi-LSTM denoising encoder is used to perform bidirectional feature extraction on the noisy multivariate time series. A spatiotemporal attention mechanism is introduced to obtain the weighted fusion of spatiotemporal attention weights. Adversarial training and Gaussian masking strategies are used for denoising. In the feature optimization layer, a sparse autoencoder is used to compress the feature space and retain key information, the features output by the Bi-LSTM denoising encoder are sparsely represented using the KL divergence penalty term, and the independence of features in each dimension is enhanced through orthogonal regularization; An inverse Bi-LSTM network is used to construct the decoder of the comprehensive calibration model. Skip connections are introduced to retain high-frequency details. The importance weights of each time step are calculated to focus on reconstructing fault-sensitive periods. Weighted time series features are used to guide the reconstruction of multi-source key parameters. The reconstructed multi-source key parameters are imported into the output layer, and a multi-task head structure is used to locate the fault type.

5. The multi-parameter comprehensive calibration method for the wet wheel abrasion tester of emulsified asphalt slurry mixture according to claim 4 is characterized in that: Import the pre-processed multi-source key parameters into the comprehensive calibration model to obtain the reconstructed data of the multi-source key parameters. Determine the fault type and mark the abnormal period based on the reconstruction error distribution, specifically: After preprocessing, the key parameters of multiple sources are standardized and time-aligned, and a multi-dimensional time series matrix is ​​constructed as the input of the comprehensive calibration model. A Bi-LSTM denoising encoder is used to extract bidirectional time series features and perform spatiotemporal attention weighting. The features output by the Bi-LSTM denoising encoder are imported into the sparse autoencoder for compression and reconstruction. The reconstructed data is used to obtain the time domain error, frequency domain error and feature space error in the bottleneck layer, generating a multi-dimensional reconstruction error. A dynamic threshold is set based on the historical normal data of the wet wheel abrasion tester of the emulsified asphalt slurry mixture, and an adaptive sliding window is set according to the short-term anomaly detection and the long-term anomaly detection. The adaptive sliding window is used to determine whether the multi-dimensional reconstruction error meets the dynamic threshold, and the abnormal period detection is performed based on the judgment result. A principal component analysis is performed based on the multi-dimensional reconstruction error, and the principal component reconstruction error is obtained as the principal component direction for principal component projection to generate a reconstruction error distribution. The fault feature pattern is matched using the Mahalanobis distance according to the abnormal period detection result and the reconstruction error distribution, and the fault category is identified according to the matching result.

6. The multi-parameter comprehensive calibration method for the wet wheel abrasion tester of emulsified asphalt slurry mixture according to claim 5 is characterized in that: The threshold value is adjusted dynamically according to the aging of the wet wheel abrasion tester for emulsified asphalt slurry mixture, specifically: Obtain the baseline normal data of the multi-source key parameters of the emulsified asphalt slurry mixture wet wheel abrasion tester, define the aging coefficient vector of the multi-source key parameters, and construct an aging model based on the baseline normal data, aging coefficient vector, and equipment usage intensity and usage time; In the aging model, a particle swarm algorithm is used to optimize the aging coefficient, the value range of the aging coefficient is initialized, and the fitness of the particle is calculated according to the deviation between the predicted normal data and the actual normal data of the current multi-source key parameters; The chaotic algorithm is introduced to perform chaotic processing on the fitness of the current particle, and the optimal position moving direction is determined by using the differential idea to update the position. The optimal solution of the aging coefficient is obtained by iterative position update until the maximum number of iterations is reached. The optimal solution of the aging coefficient is used to weight the threshold corresponding to the historical normal data and dynamically adjust the threshold.

7. The multi-parameter comprehensive calibration method for the wet wheel abrasion tester of emulsified asphalt slurry mixture according to claim 1 is characterized in that: Perform parameter collaborative calibration according to the fault type and abnormal period, record the calibration process, and verify the calibration effect, specifically: Preset a hierarchical response mechanism according to different fault types, query and set a calibration priority in the hierarchical response mechanism according to the current fault type and abnormal period, and determine the calibration time based on the calibration priority; Obtain the error profiles corresponding to the main key parameters and associated key parameters of the current fault type, introduce a stacking strategy to build a retrieval model, train a KNN model based on the distance feature vectors corresponding to Euclidean distance, Manhattan distance, and Fréchet distance, output a preset number of calibration instances, and score them based on the calibration results. Use the scoring results to determine the compensation coefficients for the main key parameters and associated key parameters. Use a preset electronic form to record the calibration process, compare the difference values ​​of the main key parameters before and after calibration, and establish a control chart to track the difference values ​​for stability monitoring and verification.

8. A multi-parameter comprehensive calibration system for a wet wheel abrasion tester for emulsified asphalt slurry mixture, characterized in that: The system includes: a memory and a processor, wherein the memory includes a multi-parameter comprehensive calibration method program for an emulsified asphalt slurry mixture wet wheel abrasion meter. When the multi-parameter comprehensive calibration method program for an emulsified asphalt slurry mixture wet wheel abrasion meter is executed by the processor, the steps of the multi-parameter comprehensive calibration method for an emulsified asphalt slurry mixture wet wheel abrasion meter as described in any one of claims 1 to 7 are implemented.

Images