Asphalt pavement compactness detection system and method based on vibration acceleration sensing technology

Abstract

The invention discloses a pavement compaction degree real-time monitoring system and method based on a high-precision vibration acceleration sensor, a signal processing algorithm and intelligent data analysis, and is suitable for vibratory roller operation. According to the system, fine vibration characteristics in the construction process are captured through the three-axis acceleration sensor installed on the vibration wheel, and the analysis precision of mechanical behaviors in the compaction process is improved by combining the viscoelastic-plastic performance, described by the three-degree-of-freedom two-dimensional viscoelastic-plastic block model, of the asphalt mixture. A compaction degree prediction model is constructed by adopting machine learning, a vibration signal is analyzed in real time, a compaction degree estimation value is output, a result is displayed through a user interface, and an optimization suggestion is provided according to a compaction effect. The system has an early warning function, gives an alarm when the compaction degree does not reach the standard, and guides constructors to adjust strategies in time, so that the road construction quality control level is effectively improved.

Description

Technical Field

[0001] The present invention relates to the field of road construction technology, and in particular to a real-time detection method and system for road compaction. The system adopts advanced vibration acceleration sensing technology and intelligent data analysis algorithm, aiming to improve construction quality and efficiency and reduce manpower dependence and detection costs. Background Art

[0002] With the acceleration of urbanization and the vigorous development of infrastructure construction, high-quality road construction has become one of the important indicators of a country's modernization level. As a core indicator for evaluating the quality of road projects, pavement compaction is directly related to the durability, bearing capacity and driving safety of roads. Traditional manual detection methods, such as nuclear density meter detection and core sampling, are not only time-consuming and labor-intensive, costly, but also difficult to achieve real-time monitoring, and are prone to missing local insufficient compaction problems, affecting overall construction efficiency and quality control.

[0003] In recent years, with the rapid development of sensor technology, Internet of Things (IoT), big data analysis and artificial intelligence (AI), intelligent and automated detection technology has gradually become the key to improving the quality of road construction. Especially in the field of vibratory compaction machinery, how to effectively use vibration acceleration sensing technology to monitor and accurately evaluate the degree of road compaction in real time has become the focus of research and development.

[0004] In the existing technology, some road compaction detection systems have begun to try to use vibration sensors to capture physical signals during the compaction process. However, these systems generally have several key problems:

[0005] 1. Poor environmental adaptability: The lack of an effective environmental factor compensation mechanism makes the sensor data susceptible to interference in a construction environment with large changes in temperature and humidity, affecting the accuracy of compaction judgment.

[0006] 2. Limited data processing capabilities: Data collection and analysis mostly rely on simple algorithms, which do not process complex and changeable vibration signals in depth, making it difficult to extract fine features that reflect the compaction state.

[0007] 3. Many existing dynamic models for vibration compaction of asphalt pavements are overly simplified and fail to fully consider the complex factors in actual working conditions. For example, the models usually assume that the asphalt pavement is uniform and isotropic, ignoring the effects of material heterogeneity and anisotropy in actual projects. Existing models often simplify the nonlinear behavior of materials. In particular, under high or low temperature conditions, the viscoelastic and plastic behaviors of asphalt mixtures are more complex, but existing models do not adequately consider this. In actual working conditions, the load of the vibratory roller changes dynamically, including factors such as periodic uplift and irregular vibration. Existing models often find it difficult to accurately simulate the effects of these dynamic loads on the asphalt pavement. Most existing models only consider single-degree-of-freedom or partial degrees of freedom in multi-degree-of-freedom systems, failing to fully reflect the multi-degree-of-freedom dynamic response in actual working conditions.

[0008] 4. Single prediction model: Modern machine learning technology, especially deep learning algorithms, has not been fully utilized. The prediction models established based on large amounts of data often have weak generalization capabilities and are not adaptable enough to different types of pavements and compaction conditions.

[0009] 5. The human-computer interaction interface is simple: the operation interface design is relatively primitive, the information display is not intuitive, and it cannot provide real-time feedback on the compaction status and effective construction guidance, which reduces work efficiency.

[0010] 6. Low system integration: The various parts of the system are relatively independent, lacking a unified management and optimization mechanism, making upgrades and maintenance difficult, and making it difficult to adapt to rapidly developing construction technology and management needs.

[0011] Chinese patent application 202011103702.4 is a real-time monitoring method for highway compaction based on artificial neural network. This method uses historical data to train neural network to establish a compaction calculation model, taking into account fill parameters, roller control parameters and sensor acquisition parameters. However, humidity is a factor that has a significant impact on road compaction. This method does not use humidity as a sensor acquisition parameter, which will cause its calculation model to be unable to accurately predict the compaction. Moreover, this method only uses the calculation model constructed by neural network, which lacks a theoretical basis and depends on the amount of data. If the amount of data is insufficient, the detection results will not be accurate enough.

[0012] Chinese patent application 202410477658.5 discloses a test method based on the acceleration of the vibrating wheel of a road roller and its correlation with the compaction degree. The method is based on an analysis of traditional compaction detection technology, factors affecting compaction, and road surface diseases caused by poor compaction quality to prevent over-compaction during the compaction process. However, the application does not involve compaction control. Compared with the present application, this method cannot adjust the compaction strategy in a timely manner during the compaction process.

[0013] Chinese patent application 202410609620.9 discloses a continuous compaction quality detection method based on BP neural network, which uses acceleration sensors and thin film pressure sensors to monitor the pressure and vibration data of the vibrating wheel during the compaction process in real time. However, this method does not install temperature and humidity sensors, and cannot obtain relevant environmental parameters in real time, which will make the real-time compaction degree prediction inaccurate. In addition, this method uses CMV as a sensor to obtain parameters. In comparison, this application uses ICMV as a sensor to obtain parameters. In actual applications, ICMV usually includes CMV and other related indicators, which can provide more refined compaction quality control.

[0014] In view of this, the present invention aims to overcome the above technical defects and proposes an asphalt pavement compaction detection system based on vibration acceleration sensing technology. The system integrates high-precision sensors, advanced data processing algorithms, powerful machine learning models, and user-friendly interactive interfaces, realizing the full chain of intelligence from data collection, processing, analysis to decision support. Through technological innovation, this system can improve the accuracy and real-time performance of pavement compaction detection, optimize the construction process, reduce manual dependence, and provide strong technical support for modern road construction. Summary of the invention

[0015] The present invention provides an asphalt pavement compaction detection system and method based on vibration acceleration sensing technology. The system overcomes many limitations of traditional compaction detection methods and, through highly integrated sensing, communication, computing and intelligent analysis technologies, achieves precise monitoring and control optimization of the pavement compaction process, thereby improving construction efficiency and quality control level.

[0016] To achieve the above object, the present invention is implemented by the following technical solutions:

[0017] An asphalt pavement compaction detection system and method based on vibration acceleration sensing technology is characterized in that a three-degree-of-freedom two-dimensional viscoelastic-plastic block model is proposed, and the Burgers model is used to describe the viscoelastic-plastic mechanical properties of asphalt mixtures. By introducing an additional degree of freedom in the asphalt surface layer, the model can simulate the periodic uplift condition of the vibrating steel wheel. The displacement response of the asphalt surface layer is determined by considering viscoelastic deformation, transient elastic deformation and viscous flow deformation. The asphalt surface layer is divided into three regional blocks according to the relative position between the asphalt surface layer and the vibrating roller. The vertical displacement of each regional block under the roller is determined by the geometric relationship to ensure that the shape of the asphalt surface is exactly the same as the shape of the roller. The asphalt surface layer is divided into blocks to simulate the change of the contact width between the vibrating steel wheel and the asphalt surface and the rolling compaction of the steel wheel, and nonlinear factors such as the periodic uplift of the vibrating steel wheel and the viscoelastic-plastic mechanical properties of the asphalt surface layer are considered. The contact force between the roller and the asphalt surface is the sum of the pressures on each block. Springs and damping are used to simulate the horizontal and vertical interactions between the blocks. The stiffness of the horizontal interaction is determined by the spring rate coefficient, while the stiffness of the vertical interaction is determined by reducing the spring rate coefficient based on the proportional relationship between the shear modulus and the elastic modulus.

[0018] The control equation of the system dynamic response is shown below.

[0019]

[0020] In the formula, m f is the mass of the roller frame; m d is the mass of the vibrating steel wheel; m a is the mass of the asphalt surface; ω is the angular velocity of the eccentric block; F e is the exciting force; the subscript i indicates the number of each block; k av , η av , k ae , η ap are the viscoelastic stiffness, elastic damping coefficient, elastic stiffness and plastic damping coefficient of the asphalt surface respectively; c f is the damping of the roller vibration reduction system; k f is the stiffness of the roller vibration reduction system; f is the roller frame displacement; z d is the displacement of the vibrating steel wheel; C1 and C2 are the integral constants generated during the integration process, which are determined by the initial viscoelastic displacement and viscoplastic displacement of the asphalt surface, respectively; F a F is the contact force between the vibrating steel wheel and the asphalt surface; x is the horizontal contact force between blocks; σ i is the stress of each block of the asphalt surface layer; G i is the horizontal shear modulus between blocks; η i is the inter-block viscosity.

[0021] A system and method for detecting compaction of asphalt pavement based on vibration acceleration sensing technology, characterized in that it comprises four subsystems: a data acquisition layer, a communication layer, a central processing unit and a user interface layer.

[0022] The data acquisition layer includes: a three-axis-in-one acceleration sensor for measuring the acceleration of the vibratory roller in the X-axis, Y-axis and Z-axis directions; a temperature sensor for measuring the real-time temperature of the asphalt mixture; a humidity sensor for measuring the humidity at the construction site; and a data acquisition module for converting the analog signals output by the acceleration sensor, temperature sensor and humidity sensor into digital signals and performing preprocessing.

[0023] The communication layer includes: LoRa wireless communication technology, which is used to transmit the data collected by the data collection layer to a central processing unit.

[0024] The central processing unit includes: a data receiving module for receiving data transmitted from the communication layer; a signal processing module for performing time-frequency domain analysis and filtering processing on the acceleration data; an intelligent analysis module for calculating the road surface compaction according to the processed acceleration data, temperature data and humidity data, and performing data analysis and evaluation; a database management module for storing the acceleration data, temperature data, humidity data and road surface compaction data, and performing data management and maintenance.

[0025] The user interface layer includes: a real-time monitoring interface for displaying the current pavement compaction, acceleration data, temperature data, and humidity data; a desktop application for providing users with access to historical data and allowing users to view and download data records; an early warning system for issuing a warning signal when the pavement compaction is lower than a predetermined threshold; and roller adjustment suggestions for providing operators with suggestions for adjusting roller working parameters based on the calculated results of the pavement compaction. The system has a modular design, allowing the sensor module, data processing module, user interface module, and other parts to be independently upgraded or replaced, and supports customized configuration according to different construction requirements. The communication protocol between modules uses the standardized protocol stack MQTT, which is convenient for the integration and expansion of third-party devices.

[0026] The above-mentioned detection system is installed on a vibratory roller. The roller mechanism includes a rear frame (1), a cover (2), an engine (3), a main clutch (4), a transfer case (5), a hydraulic pump (6), a hydraulic oil tank (7), a cab (8), a seat (9), a rearview mirror (10), a steering wheel (11), a central hinge mechanism (13), a vibrating wheel (14), and a front frame (15). The system is characterized in that a display screen (12) is installed in the cab (8) to display relevant detection data in real time. A three-axis acceleration sensor (16), a humidity sensor (17), and a temperature sensor (18) are installed on each mechanism of the roller to detect relevant data in real time. The driving speed of the vibratory roller during operation is 1.5 km / h-4 km / h.

[0027] Furthermore, the temperature sensor (18) is installed below the vibratory roller cover (2) to measure the real-time temperature of the asphalt mixture, and the humidity sensor (17) is installed above the vibratory roller rear frame (1) to monitor the humidity of the construction environment. Both the temperature sensor and the humidity sensor are high-precision sensors, wherein the temperature sensor is a digital non-contact infrared road surface temperature sensor that uses passive infrared detection technology and can measure the temperature without direct contact with the road surface; the humidity sensor is a digital temperature and humidity combination sensor with an SPI interface, with a measurement range of 0% RH to 100% RH and an accuracy of ±3% RH.

[0028] Furthermore, it comprises a three-axis acceleration sensor (16) integrating high-precision measurement functions in three axes, X, Y and Z. The sensor is installed at the wheel axle position of the vibrating wheel (14) of the vibrating roller and is used to synchronously monitor the vibration acceleration along the X-axis, Y-axis and Z-axis directions during operation. The working vibration frequency of the vibrating roller is 10 Hz-50 Hz. The sampling frequency of the vibration acceleration sensor is determined according to the Nyquist sampling theorem, and 100 times the working frequency is taken, that is, the sampling frequency of the vibration acceleration sensor is 1000 Hz-5000 Hz.

[0029] A detection system and method as described above, characterized in that it comprises a data acquisition and wireless transmission module, which comprises an analog-to-digital converter (ADC) and a microcontroller unit (MCU), wherein the ADC has a 16-bit resolution and a sampling rate of 1MSPS, and is used to convert analog signals into digital signals; the MCU integrates a 32-bit ARM Cortex-M4 processor core with an operating frequency of 120MHz, performs preliminary data filtering, noise reduction processing and compression, reduces unnecessary data transmission, and improves system response speed; LoRa technology is used to achieve long-distance, low-power data transmission, and the Advanced Encryption Standard (AES) 256-bit encryption algorithm is used during data transmission to ensure the security of transmitted data.

[0030] The above-mentioned detection system and method are characterized in that they include a central processing unit, which includes a high-performance embedded computer equipped with a quad-core ARM Cortex-A53 processor with a main frequency of 1.5GHz and a memory capacity of 2GB, and is responsible for receiving and storing all sensor data; applying advanced digital signal processing technology, including Hilbert transform, fast Fourier transform, wavelet transform and adaptive filtering, to perform time domain analysis, frequency domain analysis, time-frequency analysis, outlier detection and data smoothing on vibration data, further optimize data quality, and extract key features reflecting the compaction state of the road surface, such as peak acceleration, energy distribution, and frequency components.

[0031] A detection system and method as described above, characterized in that it comprises a machine learning model driven by big data, which utilizes an artificial neural network algorithm, and the characteristic values ​​of the neural network include fill parameters, roller control parameters and sensor acquisition parameters. Fill parameters include: gradation, mortar-sand ratio, moisture content and porosity. Roller control parameters include: paving thickness, vibration frequency, travel speed and travel direction. Sensor acquisition parameters include: compaction temperature, ambient humidity and intelligent compaction measurement value (ICMV). Regularization and cross-validation are implemented during the training process, and a compaction prediction model is constructed based on a large amount of training data.

[0032] A detection system and method as described above, characterized in that it includes a user interaction interface, which is designed using a graphical user interface (GUI) and can clearly display a real-time compaction curve, a vibration spectrum diagram, an environmental parameter display pane, a compaction suggestion area and an alarm notification bar, as well as provide historical data query and export functions; the compaction suggestion area automatically generates compaction parameter adjustment suggestions based on the prediction results, such as vibration frequency, amplitude adjustment, etc., to assist on-site decision-making; a compaction threshold is set, and when it is detected that the compaction is not up to standard, the alarm notification bar automatically alarms to prevent quality problems, and the alarm threshold can be customized by the user.

[0033] The above-mentioned detection system and method are characterized in that they include a cloud data management system, which adopts a distributed architecture, supports efficient storage and rapid retrieval of massive data, has data backup, recovery and encryption functions, and provides a RESTful API interface for integration with other management systems, supporting data synchronization and sharing.

[0034] A detection system and method as described above, characterized in that it includes an adaptive calibration module, which can automatically adjust sensor sensitivity and correction algorithm parameters after the system is initially deployed or has been in operation for a long time to ensure long-term stability and measurement accuracy. Through a built-in self-learning algorithm, the module can identify and compensate for performance degradation caused by environmental changes or mechanical wear.

[0035] The present invention has the following beneficial effects:

[0036] 1. Real-time monitoring and precise measurement: The integrated use of three-axis acceleration sensors enables the system to fully capture the vibration characteristics of the vibratory roller in the three directions of X, Y, and Z, thereby more accurately reflecting the dynamic changes in the road compaction process. The high sampling frequency ensures the integrity of data collection, and can accurately capture vibration characteristics even under high-frequency vibration, improving measurement accuracy.

[0037] 2. Brand-new dynamic model: A three-degree-of-freedom two-dimensional viscoelastic-plastic block model is proposed, and the Burgers model is used to describe the viscoelastic-plastic mechanical properties of asphalt mixtures. By introducing an additional degree of freedom in the asphalt surface layer, the model is able to simulate the periodic uplift condition of the vibrating steel wheel. The displacement response of the asphalt surface layer is determined by considering the viscoelastic deformation, transient elastic deformation, and viscous flow deformation. The change in the contact width between the vibrating steel wheel and the asphalt surface and the rolling compaction of the steel wheel are simulated by dividing the asphalt surface layer into blocks, and nonlinear factors such as the periodic uplift of the vibrating steel wheel and the viscoelastic-plastic mechanical properties of the asphalt surface layer are considered.

[0038] 3. Comprehensive construction environment monitoring: The installation of temperature sensors and humidity sensors can monitor the temperature and humidity conditions of the construction environment in real time, which is essential for evaluating construction quality and optimizing the construction process. Multiple sensors work together, combined with acceleration sensors, to form a comprehensive monitoring network that can better understand the various variables and their interactions in the construction process.

[0039] 4. Advanced data analysis and modeling: The central processing unit uses advanced signal processing technology to effectively extract key features and conduct in-depth analysis to help identify potential problems in the compaction process. The application of machine learning models, especially artificial neural network algorithms, can be trained based on large amounts of data to achieve accurate prediction of compaction.

[0040] 5. Adaptive calibration and modular design: The adaptive calibration module ensures the long-term stable operation of the system and reduces maintenance costs. The modular design allows the system to be flexibly configured according to different construction requirements, improving the adaptability and scalability of the system.

[0041] In summary, the present invention provides a highly integrated and intelligent road compaction detection system, which significantly improves the quality control level of compaction operations, reduces construction costs, and can meet the personalized needs of different engineering projects. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] Figure 1 This is a dynamic model diagram of the asphalt surface layer vibration compaction of the present invention, in which: m f is the mass of the roller frame; md is the mass of the vibrating steel wheel; m a is the mass of the asphalt surface; ω is the angular velocity of the eccentric block; F e is the exciting force; the subscript i indicates the number of each block; k av , η av , k ae , η ap are the viscoelastic stiffness, elastic damping coefficient, elastic stiffness and plastic damping coefficient of the asphalt surface respectively; c f is the damping of the roller vibration reduction system; k f is the stiffness of the roller vibration reduction system; f is the roller frame displacement; z d is the displacement of the vibrating steel wheel; G i is the horizontal shear modulus between blocks; η i is the inter-block viscosity.

[0043] Figure 2 is a system architecture diagram of the present invention;

[0044] Figure 3 The data processing and intelligent analysis flow chart of the present invention;

[0045] Figure 4 The present invention is a schematic diagram of the layout of the vibration acceleration sensor and the environmental sensor, in which: frame 1, cover 2, engine 3, main clutch 4, transfer case 5, hydraulic pump 6, hydraulic oil tank 7, cab 8, seat 9, rearview mirror 10, steering wheel 11, display screen 12, central hinge mechanism 13, vibration wheel 14, front frame 15, three-axis acceleration sensor 16, humidity sensor 17, temperature sensor 18;

[0046] Figure 5 This is a neural network structure diagram of the present invention. DETAILED DESCRIPTION

[0047] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0048] In a newly built expressway project with a total length of 100 kilometers, in order to ensure that the compaction of the road base and surface layers meets the design requirements, the real-time detection system of the present invention is used to replace the traditional random sampling and laboratory testing methods.

[0049] The system architecture of the present invention is as follows Figure 2 shown.

[0050] Data acquisition module: installed on the roller, integrated with analog-to-digital converter (ADC) and microcontroller, responsible for the preliminary processing and local storage of sensor data. Wireless transmission module: using LoRa wireless communication technology, each sensor node transmits data to the central processing unit located in the construction site command center through a low-power wireless module. The center is equipped with a high-performance server for centralized data processing and storage. Data receiving module: receives sensor data from the roller and performs preliminary sorting and formatting. Signal processing module: using digital signal processing technology, low-pass filter removes high-frequency noise, adaptive filtering improves data quality, identifies and eliminates abnormal data points caused by mechanical failure or external interference, and extracts key vibration characteristics. Intelligent analysis module: based on machine learning models, analyzes feature data, predicts compaction degree, and provides compaction strategy recommendations. Database management module: stores historical data and supports data query, backup and recovery. Real-time monitoring interface: displays vibration parameter trends, environmental parameters, compaction strategy recommendations, etc., for operators to monitor and adjust in real time. Desktop application: provides a convenient access portal for on-site engineers and managers, supporting real-time data viewing, alarm reception, and historical data analysis functions. Early warning system: When the compaction degree does not meet the standard or the predicted result deviates from the expected, the alarm is automatically triggered. Compaction strategy adjustment suggestions: According to the compaction degree prediction results, it provides operational suggestions such as adjusting the vibration frequency and driving speed to optimize the compaction effect. It also recommends adjusting parameters such as vibration frequency, amplitude or compaction path. The operator can make real-time adjustments based on the suggestions to ensure that the road surface is compacted evenly and meets the design standards.

[0051] Data processing and intelligent analysis process such as Figure 3 As shown, it starts with sensor data collection, goes through signal filtering, noise reduction, and feature extraction. The machine learning algorithm area performs intelligent analysis and decision-making, uses artificial neural networks, and predicts compaction based on a large amount of historical data, and further connects the early warning system and roller control adjustment.

[0052] Sensor deployment: A large vibratory roller is equipped with a three-axis acceleration sensor located at the center of the vibrating wheel axle to ensure full capture of vibration characteristics. At the same time, the roller is also equipped with two environmental sensors (one for temperature and one for humidity). The temperature sensor is the HQC-WR2 digital non-contact infrared road surface temperature sensor, and the humidity sensor is the SHT31, which are installed in different parts of the roller, such as Figure 5 As shown, to monitor the working environment.

[0053] System initialization and calibration: Before construction, all sensors were calibrated to ensure the accuracy of data collection.

[0054] When collecting general signals, the sampling frequency needs to be set to prevent distortion of the sampling frequency. For example, in this example, the sampling frequency of the roller vibration acceleration signal is set to more than 100 times the vibration frequency to prevent the sampling frequency from being too low, which will cause sampling distortion. Therefore, three sampling frequencies of 1000Hz, 2000Hz, and 5000Hz are used to collect data for the middle and lower layers of the asphalt pavement.

[0055] First, set the acquisition frequency to 1000Hz, the measurement type to acceleration, the measurement type to voltage, and the input mode to ICP. After setting all the parameters, debug the equipment to ensure that all equipment is working properly and perform offline sampling. Communicate with the roller driver to keep the roller rolling speed at around 3.6km / h and only start the front wheel vibration when moving forward. In order to prevent the temperature from dropping below the reasonable rolling temperature, the roller starts rolling after about 20 meters of paving, rolling a section of road five times.

[0056] After this sampling frequency ends, the sampling frequency is reset to 2000 Hz and 5000 Hz, and sampling is performed on different sections of the same road surface according to the above method.

[0057] Synchronization mechanism: Ensure that data collection from all sensors is synchronized to facilitate consistency in subsequent data analysis.

[0058] Building a neural network framework: Figure 5 As shown in the figure, the eigenvalues ​​of the neural network include the parameters of the compacted body, the control parameters of the roller and the parameters acquired by the sensor. The parameters of the compacted body include: gradation, mortar-sand ratio, moisture content and porosity. The control parameters of the roller include: paving thickness, vibration frequency, driving speed and driving direction. The parameters acquired by the sensor include: compaction temperature, ambient humidity and intelligent compaction measurement value (ICMV). The target value is the degree of compaction. A compaction quality evaluation model based on an artificial neural network is established, with one layer in the middle layer, and the neurons between layers are fully connected. There is no connection between the neurons in the layer. The input layer contains 2000 neurons and the output layer contains 3 neurons.

[0059] Training the neural network: After obtaining the available data set, use the multi-layer validation method to divide the training set, test set, and validation set. The multi-layer validation method means that the validation set contains multiple data subsets, and each data subset needs to contain data combinations with different eigenvalues. Each data subset is a series of data combinations with only one eigenvalue different and the same other eigenvalues, or each data subset is a series of data combinations with different eigenvalues ​​and the same eigenvalues ​​of other classes. For example, the data of subset one is the data of the same section of highway with the same fill parameters but different roller setting parameters, the data of subset two is the data of roller parameters with the same roller but rolling fills of different properties, and subset three is the data of different sections of highway, etc.

[0060] The MATLAB platform has powerful numerical computing capabilities. With its rich function library, it is widely used by the engineering community to process simulation calculations for various engineering applications. Based on the MATLAB platform, the nnstart function is used to call the artificial neural network development interface. The Patern Recognition app module is selected to model and train the neural network compaction quality evaluation model. The network structure is as follows: Figure 4 shown.

[0061] The training sample set is randomly divided into three independent subsets: training set, validation set and test set, which account for 15%, 15% and 70% of the training sample capacity respectively. The training set is used to train the network, that is, to calculate gradients and update weights. The validation set does not participate in the training, but the validation set is used to verify the error (loss function value) at each step of the training iteration. At the beginning of the training, the errors of the training set and the validation set usually decrease with the iteration. As the iteration proceeds, if overfitting occurs, the error of the training set will continue to decrease while the error of the validation set will increase. In this way, the validation set can be used to determine whether overfitting occurs. The test set does not participate in the training. After the training is completed, the trained network is tested with the test set.

[0062] The feature vectors are analyzed using a trained machine learning model (neural network) to learn the relationship between compaction and vibration characteristics under different conditions.

[0063] Performance recording and analysis: Record the data of each compaction operation to facilitate later analysis and optimization strategy formulation.

[0064] Effect evaluation: After the construction was completed, the accuracy of the real-time detection system was verified by randomly sampling and drilling core tests on the compacted road sections. The results showed that the average error between the compaction degree predicted by this system and the laboratory test results was less than 2%, which was significantly better than the traditional detection method.

[0065] In this embodiment, the asphalt pavement compaction detection system based on vibration acceleration sensing technology demonstrates its significant advantages in improving construction efficiency and ensuring construction quality. Through real-time monitoring and intelligent adjustment, it effectively avoids pavement quality problems caused by uneven compaction, reduces unnecessary repeated operations, improves overall construction efficiency, and provides a set of efficient and reliable intelligent solutions for highway construction.

[0066] Although the present invention has been described in detail above by general description and specific embodiments, it is obvious to those skilled in the art that some modifications or improvements can be made to the present invention. Therefore, these modifications or improvements made without departing from the spirit of the present invention all belong to the scope of protection claimed by the present invention.

Claims

1. A system and method for detecting compaction of asphalt pavement based on vibration acceleration sensing technology, characterized in that: A three-degree-of-freedom two-dimensional viscoelastic-plastic block model is proposed, which takes into account nonlinear factors such as the periodic uplift of the vibrating steel wheel and the viscoelastic-plastic mechanical properties of the asphalt surface layer. The control equation of the system dynamic response is shown below. In the formula, m f is the mass of the roller frame; m d is the mass of the vibrating steel wheel; m a is the mass of the asphalt surface; ω is the angular velocity of the eccentric block; F e is the exciting force; the subscript i indicates the number of each block; k av , η av , k ae , η ap are the viscoelastic stiffness, elastic damping coefficient, elastic stiffness and plastic damping coefficient of the asphalt surface respectively; c f is the damping of the roller vibration reduction system; k f is the stiffness of the roller vibration reduction system; f is the roller frame displacement; z d is the displacement of the vibrating steel wheel; C1 and C2 are the integral constants generated during the integration process, which are determined by the initial viscoelastic displacement and viscoplastic displacement of the asphalt surface, respectively; F a F is the contact force between the vibrating steel wheel and the asphalt surface; x is the horizontal contact force between blocks; σ i is the stress of each block of the asphalt surface layer; G i is the horizontal shear modulus between blocks; η i is the inter-block viscosity.

2. A system and method for detecting compaction of asphalt pavement based on vibration acceleration sensing technology, characterized in that: It includes four subsystems: data acquisition layer, communication layer, central processing unit and user interface layer. The data acquisition layer includes: a three-axis-in-one acceleration sensor, a temperature sensor, and a humidity sensor; and a data acquisition module that converts analog signals output by the acceleration sensor, the temperature sensor, and the humidity sensor into digital signals and performs preprocessing. The communication layer adopts LoRa wireless communication technology to transmit the data collected by the data collection layer to the central processing unit. The central processing unit includes: a data receiving module, a signal processing module, an intelligent analysis module, and a database management module. The user interface layer includes: a real-time monitoring interface, which includes the current road surface compaction, acceleration data, temperature data and humidity data; a desktop application, which includes the function of users accessing historical data and allows users to view and download data records; an early warning system, which sends a warning signal when the road surface compaction is lower than a predetermined threshold; and roller adjustment suggestions, which provide operators with suggestions for adjusting the roller working parameters based on the calculation results of the road surface compaction. The system has a modular design, allowing the sensor module, data processing module, user interface module and other parts to be independently upgraded or replaced, and the communication protocol between modules uses the standardized protocol stack MQTT.

3. A system and method according to claim 2, installed on a vibratory roller, the roller mechanism comprising a rear frame (1), a cover (2), an engine (3), a main clutch (4), a transfer case (5), a hydraulic pump (6), a hydraulic oil tank (7), a cab (8), a seat (9), a rearview mirror (10), a steering wheel (11), a central hinge mechanism (13), a vibrating wheel (14), and a front frame (15), characterized in that A display screen (12) is installed in the cab (8) to display relevant detection data in real time. A three-axis acceleration sensor (16), a humidity sensor (17) and a temperature sensor (18) are installed on each mechanism of the roller to detect relevant data in real time. The driving speed of the vibratory roller during operation is 1.5 km / h-4 km / h.

4. A system and method according to claim 2, characterized in that: The invention comprises a three-axis acceleration sensor (16) which integrates high-precision measurement functions of three axes, namely, X, Y and Z. The sensor is installed at the wheel axle position of the vibrating wheel (14) of the vibrating roller to synchronously monitor the vibration acceleration along the X-axis, Y-axis and Z-axis directions during operation. The working vibration frequency of the vibrating roller is 10 Hz-50 Hz. The sampling frequency of the vibration acceleration sensor is determined according to the Nyquist sampling theorem, and 100 times the working frequency is taken, that is, the sampling frequency of the vibration acceleration sensor is 1000 Hz-5000 Hz.

5. A system and method according to claim 2, characterized in that: The invention comprises a temperature sensor (18) and a humidity sensor (17), wherein the temperature sensor (18) is installed below the cover (2) of the vibratory roller to measure the real-time temperature of the asphalt mixture, and the humidity sensor (17) is installed above the rear frame (1) of the vibratory roller to monitor the humidity of the construction environment, wherein the temperature sensor is a digital non-contact infrared road surface temperature sensor, which adopts passive infrared detection technology; and the humidity sensor is a digital temperature and humidity combination sensor with an SPI interface, which has a measurement range of 0%RH to 100%RH and an accuracy of ±3%RH.

6. A system and method according to claim 2, characterized in that: It includes a data acquisition and wireless transmission module, which includes an analog-to-digital converter (ADC) and a microcontroller unit (MCU). The ADC has a 16-bit resolution and a sampling rate of 1MSPS, and is used to convert analog signals into digital signals; the MCU integrates a 32-bit ARM Cortex-M4 processor core with an operating frequency of 120MHz to perform preliminary data filtering, noise reduction processing and compression; it adopts LoRa technology, and uses the Advanced Encryption Standard (AES) 256-bit encryption algorithm during data transmission.

7. A system and method according to claim 2, characterized in that: It includes a central processing unit, which includes a high-performance embedded computer equipped with a quad-core ARM Cortex-A53 processor with a main frequency of 1.5GHz and a memory capacity of 2GB. It is responsible for receiving and storing all sensor data. It applies advanced digital signal processing techniques, including Hilbert transform, fast Fourier transform, wavelet transform and adaptive filtering, to perform time domain analysis, frequency domain analysis, outlier detection and data smoothing on vibration data, further optimize data quality, and extract key features reflecting the compaction status of the road surface, such as peak acceleration, energy distribution, and frequency components.

8. A system and method according to claim 2, characterized in that: It includes a machine learning model driven by big data. The model uses an artificial neural network algorithm. The eigenvalues ​​of the neural network include fill parameters, roller control parameters, and sensor acquisition parameters. Fill parameters include: gradation, mortar-sand ratio, moisture content, and porosity. Roller control parameters include: paving thickness, vibration frequency, travel speed, and travel direction. Sensor acquisition parameters include: compaction temperature, ambient humidity, and intelligent compaction measurement value (ICMV). Regularization and cross-validation are implemented during the training process, and a compaction prediction model is built based on a large amount of training data.

9. A system and method according to claim 2, characterized in that: It includes a user interaction interface designed with a graphical user interface (GUI), which includes a compaction curve graph, a vibration spectrum graph, an environmental parameter display pane, a compaction suggestion area and an alarm notification bar, and provides historical data query and export functions; in the compaction suggestion area, based on the prediction results, the system automatically generates compaction parameter adjustment suggestions, such as vibration frequency and amplitude adjustment, to assist on-site decision-making; set the compaction threshold, and when it is detected that the compaction is not up to standard, the alarm notification bar automatically alarms to prevent quality problems. The alarm threshold can be customized by the user.

10. A system and method according to any one of claims 1 to 9, characterized in that: Contains adaptive calibration module with built-in self-learning algorithm.

Images