Roadbed compaction detection system

Abstract

The invention relates to a roadbed compaction detection system. The system comprises a high-precision positioning module, a data acquisition module, an edge calculation module, an intelligent display module and a cloud collaboration platform. The high-precision positioning module acquires centimeter-level positioning data of the road roller in real time; the data acquisition module acquires a vertical vibration signal; the edge calculation module synchronously processes the positioning data and the vibration signal, generates quality data of a compaction track, compaction times, loose laying thickness and compaction thickness, and generates alarm information; the intelligent display module maps the compaction quality data to construction display equipment through a thermodynamic diagram; and the cloud collaboration platform integrates multi-device collaboration operation data, and realizes quality tracing based on a block chain technology. According to the invention, through real-time data processing and multi-module cooperative work, high-efficiency and accurate detection and quality control of roadbed pressing construction are realized.

Description

Technical Field

[0001] The invention belongs to the field of roadbed compaction detection, and in particular relates to a roadbed compaction detection system. Background Art

[0002] In the field of roads and civil engineering, roadbed compaction is a key link to ensure the stability and durability of road structures. Traditional roadbed compaction detection methods mainly rely on manual detection and empirical judgment, and have problems such as low detection accuracy, low efficiency, and incomplete data records. For example, commonly used detection methods such as the ring knife method and sand injection method are not only time-consuming and labor-intensive, but also difficult to achieve real-time detection and data traceability.

[0003] With the development of the Internet of Things, big data and artificial intelligence technologies, automated and intelligent detection systems are gradually being used in roadbed compaction construction. However, the existing detection systems still have some shortcomings. For example, some systems lack high-precision positioning data, resulting in inaccurate statistics of compaction trajectories and compaction passes; in addition, the data integration and traceability functions are not perfect enough to meet the needs of modern construction management.

[0004] Therefore, it is of great practical significance and application value to develop a high-precision, high-efficiency and intelligent roadbed compaction detection system to realize real-time monitoring, data analysis and quality traceability of compaction construction. Summary of the invention

[0005] Based on this, it is necessary to provide a roadbed compaction detection system to address the above technical issues and solve the problems existing in the prior art.

[0006] In the first aspect, the present application provides a roadbed compaction detection system, including a high-precision positioning module, a data acquisition module, an edge computing module, an intelligent display module and a cloud collaboration platform:

[0007] High-precision positioning module, used to obtain centimeter-level positioning data of the roller in real time;

[0008] A data acquisition module is used to collect vertical vibration signals generated by the roller during the roadbed compaction process;

[0009] The edge computing module is used to generate quality data of compaction trajectory, compaction passes, loose paving thickness and compaction thickness in real time based on positioning data and vibration signals;

[0010] Intelligent display module, used to map compaction quality data to construction display equipment through thermal maps;

[0011] A cloud-based collaborative platform that integrates data from collaborative operations of multiple devices and enables quality traceability based on blockchain technology.

[0012] In one of the embodiments, the edge computing module includes a trajectory generation unit, a pass calculation unit, and a thickness modeling unit:

[0013] A trajectory generation unit, used for generating a compaction trajectory using a B-spline curve to fit the positioning data;

[0014] A pass calculation unit, used to generate the compaction pass number by dividing the working interval into preset sizes and counting the number of times the compaction track covers each interval;

[0015] The thickness modeling unit is used to establish a relationship model between loose paving thickness and compacted thickness based on the characteristic parameters of the vibration signal and experimental data.

[0016] In one embodiment, the compacted thickness is calculated by the following method:

[0017] Use the following formula to calculate the autocorrelation function R(k) corresponding to the vibration response signal:

[0018]

[0019] Where x[n] represents the vibration response signal, N represents the signal length, and k represents the lag time;

[0020] Use the following formula to identify the fundamental frequency f of the vibration response signal x[n] 0 :

[0021]

[0022] Among them, k p represents the hysteresis of the first peak of R(k);

[0023] Calculate the Fourier transform X(f) of the vibration response signal x[n] using the following formula:

[0024]

[0025] Among them, e represents a natural number, f represents a frequency, and j represents an imaginary unit;

[0026] Identify the fundamental frequency f 0 and its harmonics f n =nf 0 , where n = 1, 2, ...;

[0027] Use the following formula to calculate the amplitude A of the identified fundamental frequency and harmonics 基 and the amplitude of the nth harmonic A 谐n :

[0028] A 基 =|X(f 0 )|,A 谐n =|X(nf0 )|;

[0029] Calculate the compaction thickness H using the following formula:

[0030]

[0031] Among them, v s represents the speed of the roller, q represents the quality index of the roadbed, and a, b and c represent constants determined by experimental calibration.

[0032] In one embodiment, an experiment adjustment module is also included, and the experiment adjustment module is used to:

[0033] The experimental area is divided into qualified area and unqualified area along the center line. The width of both areas is 5 meters. The loose paving thickness of the qualified area is x according to the maximum particle size D of the material. m and design target compaction H d Sure:

[0034] x=0.4·D m +0.2 H d

[0035] Among them, the value range of x is 15cm to 35cm;

[0036] The loose paving thickness of the unqualified area is y=x+Δh, where Δh is determined by the construction standard;

[0037] Set the number of rolling times w in the qualified area, w is based on the target compaction degree K t calculate:

[0038]

[0039] Where, μ represents the single rolling efficiency coefficient, μ = 0.15, K t ≥95%;

[0040] The number of times the unqualified area is rolled is z=w-Δj, where Δj is the preset number of deviations;

[0041] Calculate the quality data deviation of compaction trajectory, compaction passes, loose paving thickness and compaction thickness detected by the roller and manually, and obtain the detection deviation information;

[0042] According to the detection deviation information, the least squares method is used to adjust the calculation parameters of the edge computing module.

[0043] In one embodiment, the edge computing module further includes an elevation analysis unit, and the elevation analysis unit is used to:

[0044] Based on the real-time positioning data of adjacent working areas, determine the relationship between layers:

[0045] If the difference in elevation between the current working interval and the adjacent working interval is ≥12cm, the working interval with a higher elevation is defined as the new layer, and the working interval with a lower elevation is defined as the old layer;

[0046] If the elevation difference between the current working area and the adjacent working area is less than 12 cm, they are considered to be on the same floor;

[0047] Generate inter-layer inspection reports based on working intervals and inter-layer relationships.

[0048] In one embodiment, the edge computing module is further used to:

[0049] When a new layer is rolled for the first time, the loose paving thickness is detected in real time. When it is detected that the loose paving thickness does not meet the preset value range ΔH, a loose paving alarm message is generated. The preset value range ΔH is determined by the construction standard.

[0050] When the compaction construction of the new or old layer is completed, a compaction test report is generated in combination with historical data;

[0051] Based on preset conditions, when the compaction detection report does not meet the preset requirements, a compaction alarm message is generated.

[0052] In one embodiment, the working method of the intelligent display module includes:

[0053] Get the thermal map color mapping rules corresponding to the loose paving thickness, compaction passes, and compaction thickness indicators;

[0054] Based on the color mapping rules of the thermal map, and according to the positioning data and the compaction trajectory, compaction passes and loose paving thickness information, image rendering is performed to generate a thermal trajectory map of loose paving thickness, compaction passes and compaction thickness;

[0055] The thermal trajectory diagram of compaction quality information is generated by integrating the thermal trajectory diagrams of compaction passes, loose paving thickness and compaction thickness using thermal map overlay technology;

[0056] Utilizing SLAM technology, the image on the construction display device is dynamically updated based on the thermal trajectory map of compaction quality information.

[0057] In one embodiment, the cloud collaboration platform is also used to:

[0058] Obtain the operating data of multiple rollers and generate a full-section roadbed compaction 3D surface model based on the operating data;

[0059] Iteratively optimize the feature parameter library based on federated machine learning technology, which contains data features and model training parameters;

[0060] The hash value of the compaction quality data is written into the Ethereum smart contract to generate an electronic construction log.

[0061] In one embodiment, a self-learning optimization module is included, which is used to:

[0062] Obtain matching data sets of vibration signals and manual detection results under different working conditions to form a knowledge base;

[0063] Use the knowledge base to train machine learning models;

[0064] The Bayesian optimization algorithm is used to dynamically adjust the feature extraction threshold of the vibration signal and the weight parameters of the learning model;

[0065] Based on field data, the learning model is used to generate prediction information and calculate system confidence;

[0066] When the system confidence is less than the preset value, the on-site calibration process is triggered and the knowledge base is updated.

[0067] In one embodiment, the detection report includes the warning number, event level, warning type, warning message and timestamp:

[0068] The warning pile number is used to identify the specific location where the warning occurs;

[0069] The event level is used to classify warning events into three levels: severe, moderate, and minor according to the severity of the warning;

[0070] Warning type is used to specify the type of warning, including unqualified compaction passes and unqualified compaction thickness;

[0071] Warning messages are used to describe the specific content of the warning, including the cause of the warning and possible impact.

[0072] The roadbed compaction detection system provided by this application uses a high-precision positioning module to capture the roller position in real time, a data acquisition module to obtain the vertical vibration spectrum, and an edge computing module to dynamically generate compaction trajectories, pass statistics and thickness indicators. The intelligent display module generates a heat map to intuitively present the quality deviation area. At the same time, it relies on a cloud-based collaborative platform to build tamper-proof construction files and form a full-chain digital closed-loop control system. It solves the technical pain points of traditional manual detection, such as large trajectory errors, lagging thickness detection, low data credibility, and difficulty in multi-device collaboration. BRIEF DESCRIPTION OF THE DRAWINGS

[0073] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the related technologies, the drawings required for use in the embodiments or the related technical descriptions are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0074] Figure 1It is a structural schematic diagram of a roadbed compaction detection system of the present invention;

[0075] Figure 2 It is a structural schematic diagram of the edge computing module of the present invention; DETAILED DESCRIPTION

[0076] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0077] The present application discloses a roadbed compaction detection system, which is mainly suitable for road construction, especially roads that require intelligent monitoring of roadbed compaction quality. Application scenarios include highway or railway roadbed construction, municipal road reconstruction and expansion, and special foundation treatment.

[0078] In one embodiment, Figure 1 As shown, a roadbed compaction detection system is provided. This embodiment takes the method integrated in the roadbed ballast terminal as an example. It can be understood that the system can also be deployed on a server, and realize two-way transmission of data and instructions through the interaction between the terminal and the server. In this embodiment, the system includes a high-precision positioning module 101, a data acquisition module 102, an edge computing module 103, an intelligent display module 104 and a cloud collaboration platform 105:

[0079] High-precision positioning module, used to obtain centimeter-level positioning data of the roller in real time;

[0080] Specifically, the high-precision positioning module adopts carrier phase differential technology, and realizes centimeter-level real-time positioning of the three-dimensional coordinates of the roller through the joint solution of GNSS (Global Navigation Satellite System) satellite signals and ground reference stations. For example, a Beidou satellite CORS (Continuous Operational Reference System) station is established to perform real-time differential calculations on the location information within the service range, realize centimeter-level positioning of the roller's rolling trajectory, and provide a spatial reference for trajectory generation.

[0081] The data acquisition module 102 is used to collect vertical vibration signals generated by the roller during the roadbed compaction process;

[0082] Exemplarily, the data acquisition module 102 can collect vertical vibration signals during construction through an acceleration sensor installed on the roller. For example, through a high-precision three-axis acceleration sensor, the vertical vibration signals during construction are collected in real time at a sampling rate of 2000 Hz to provide a calculation data basis for the edge computing module 103.

[0083] The edge computing module 103 is used to generate quality data of compaction trajectory, compaction passes, loose paving thickness and compaction thickness in real time according to the positioning data and vibration signals;

[0084] Specifically, the edge computing module 103 generates compaction trajectory, compaction passes, loose laying thickness and quality data of compaction thickness by real-time fusion of centimeter-level positioning data and spectral characteristics of vibration signals. For example, based on the positioning data and vibration signal data, high-precision compaction trajectory data is generated according to an adaptive B-spline curve (a curve representation method in mathematics and computer graphics) fitting algorithm, and compaction pass data of 0.3m×0.3m units is statistically generated according to dynamic grid division technology. A thickness prediction model is constructed by combining the fundamental frequency attenuation rate of the vibration signal and the harmonic energy ratio, and quality data of loose laying thickness and compaction thickness are generated according to the thickness prediction model.

[0085] Intelligent display module 104, used to map compaction quality data to a construction display device through a thermal map;

[0086] Exemplarily, the intelligent display module 104 may use color space encoding technology to generate a compaction quality data heat map, implement augmented reality overlay through the ORB-SLAM2 algorithm, and dynamically render a three-dimensional heat map on the roller cockpit display screen.

[0087] The cloud collaboration platform 105 is used to integrate the collaborative operation data of multiple devices and realize quality traceability based on blockchain technology.

[0088] Specifically, the cloud-based collaborative platform 105 can build a consortium chain network based on the Hyper Ledger Fabric (enterprise-level licensed distributed ledger technology) architecture, design smart contracts to implement chain storage of hash values ​​of multi-device data packets, support global model optimization under the federated learning framework, and generate verifiable electronic construction logs, forming a full life cycle quality traceability system from single-machine operation to multi-construction site linkage.

[0089] The above-mentioned roadbed compaction detection system works together through the high-precision positioning module, data acquisition module 102, edge computing module 103, intelligent display module 104 and cloud collaboration platform 105. The high-precision positioning module obtains the centimeter-level positioning data of the roller in real time, and the data acquisition module 102 synchronously collects the vertical vibration signal generated by the roller during operation. The edge computing module 103 receives data from the high-precision positioning module and the data acquisition module 102, and generates quality data such as compaction trajectory, compaction passes, loose paving thickness and compaction thickness in real time, and generates alarm information when an abnormality is detected. The intelligent display module 104 intuitively maps these quality data to the construction display device in the form of a heat map, so that construction personnel can understand the compaction situation in real time. The cloud collaboration platform 105 integrates the collaborative operation data of multiple devices, and uses blockchain technology to achieve data immutability and quality traceability. It solves the problems of low detection accuracy, low efficiency, incomplete data records caused by traditional roadbed compaction detection methods relying on manual detection and experience judgment, and the lack of data integration and imperfect traceability functions of the existing detection system. It provides a high-precision, high-efficiency and intelligent roadbed compaction detection system, which realizes real-time monitoring, data analysis and quality traceability of compaction construction, and has important practical significance and application value.

[0090] In one embodiment, the edge computing module 103 includes a trajectory generation unit, a pass calculation unit, and a thickness modeling unit:

[0091] A trajectory generation unit, used for generating a compaction trajectory using a B-spline curve to fit the positioning data;

[0092] A pass calculation unit, used to generate the compaction pass number by dividing the working interval into preset sizes and counting the number of times the compaction track covers each interval;

[0093] The thickness modeling unit is used to establish a relationship model between loose paving thickness and compacted thickness based on the characteristic parameters of the vibration signal and experimental data.

[0094] Exemplarily, the edge computing module 103 may be composed of a trajectory generation unit, a pass calculation unit, and a thickness modeling unit. Among them, the trajectory generation unit adopts an adaptive B-spline curve fitting technology, which may be a high-precision compaction trajectory with continuous curvature through dynamic elimination of abnormal positioning points and control point encryption strategies, thereby improving the accuracy and reliability of the trajectory. The pass calculation unit may be based on quadtree dynamic grid division and vibration energy triggering mechanism, combined with a time window overlap detection algorithm to obtain the number of rolling times, thereby improving the accuracy and efficiency of pass statistics. The thickness modeling unit may be an exponential-logarithmic hybrid prediction model constructed by extracting the fundamental frequency attenuation rate and harmonic energy ratio characteristics of the vibration signal, and automatically updating the coefficient online every 50m of rolling, thereby realizing high-precision real-time calculation of loose paving thickness and compaction thickness. It provides a reliable quality control means for roadbed compaction construction.

[0095] In one embodiment, the compacted thickness is calculated by the following method:

[0096] S110. Calculate the autocorrelation function R(k) corresponding to the vibration response signal using the following formula:

[0097]

[0098] Where x[n] represents the vibration response signal, N represents the signal length, and k represents the lag time;

[0099] S120, using the following formula, identify the fundamental frequency f of the vibration response signal x[n] 0 :

[0100]

[0101] Among them, k p represents the hysteresis of the first peak of R(k);

[0102] S130. Calculate the Fourier transform X(f) of the vibration response signal x[n] using the following formula:

[0103]

[0104] Among them, e represents a natural number, f represents a frequency, and i represents an imaginary unit;

[0105] S140, identify the fundamental frequency f 0 and its harmonics f n =nf 0 , where n = 1, 2, above;

[0106] S150, using the following formula, calculate the amplitude A of the identified fundamental frequency and harmonics 基 and the amplitude of the nth harmonic A 谐n :

[0107] A 基 =|X(f 0 )|,A 谐n =|X(nf 0 )|;

[0108] S160. Calculate the compaction thickness H using the following formula:

[0109]

[0110] Among them, v s represents the speed of the roller, q represents the quality index of the roadbed, and a, b and c represent constants determined by experimental calibration.

[0111] Exemplarily, the formula calculates the autocorrelation function of the vibration response signal through signal autocorrelation analysis, uses discrete Fourier transform to extract frequency domain features of the signal, obtains thickness prediction modeling based on multiple parameters, realizes efficient thickness detection, and has the characteristics of multi-parameter fusion and flexible engineering calibration.

[0112] In one embodiment, an experiment adjustment module is further included, and the experiment adjustment module is used to:

[0113] S210, the experimental area is divided into qualified area and unqualified area along the center line, the width of the two areas is 5 meters, the loose paving thickness of the qualified area is x according to the maximum particle size D of the material m and design target compaction H d Sure:

[0114] x=0.4·D m +0.2 H d

[0115] Wherein, the value range of x is 15cm to 35cm, and the loose paving thickness of the unqualified area is y=x+Δh, where Δh is determined by the construction standard;

[0116] S220, set the number of rolling times w in the qualified area, w is based on the target compaction degree K t calculate:

[0117]

[0118] Where, μ represents the single rolling efficiency coefficient, μ = 0.15, K t ≥95%;

[0119] The number of times the unqualified area is rolled is z=w-Δj, where Δj is the preset number of deviations;

[0120] S230, calculating the quality data deviation of compaction trajectory, compaction passes, loose paving thickness and compaction thickness detected by the roller and manually, and obtaining detection deviation information;

[0121] According to the detection deviation information, the calculation parameters of the edge calculation module 103 are adjusted using the least squares method.

[0122] Specifically, the experimental adjustment module optimizes the system parameters through a dual-zone comparison experiment, which can be done by dividing the area into qualified and unqualified areas, and setting different roadbed compaction conditions for the qualified and unqualified areas respectively. By comparing the equipment detection data with the manual measurement results, the trajectory coincidence deviation, compaction deviation and thickness deviation are calculated, and a least squares optimization model is constructed based on this. The optimization model is iteratively solved, and the optimization parameters are output to realize adaptive calibration of the edge computing module 103, which can ensure the detection accuracy and reliability of the system under different working conditions.

[0123] In one embodiment, the edge computing module 103 further includes an elevation analysis unit, which is used to:

[0124] S310, judging the relationship between layers based on the real-time positioning data of adjacent working intervals:

[0125] If the difference in elevation between the current working interval and the adjacent working interval is ≥12cm, the working interval with a higher elevation is defined as the new layer, and the working interval with a lower elevation is defined as the old layer;

[0126] If the elevation difference between the current working area and the adjacent working area is less than 12 cm, they are considered to be on the same floor;

[0127] S320, generating an inter-layer detection report according to the working interval and the inter-layer relationship.

[0128] Specifically, identifying new layers and old layers is the basis for detecting loose laying thickness, compaction times, and compaction thickness. The elevation analysis unit in the edge computing module 103 determines the inter-layer relationship based on the real-time positioning data of adjacent working intervals. When the elevation difference of adjacent grids is detected to be ≥12cm, the inter-layer judgment logic is triggered, and the area with sudden elevation changes is automatically marked as the boundary between the new and old layers. The adjacent working intervals are judged as new layers or old layers according to the height of the elevation. If the elevation difference is less than 12cm, they are judged to be the same layer, and an inter-layer detection report is generated based on this information, which provides accurate layering information for roadbed construction and improves construction accuracy and efficiency.

[0129] In one embodiment, the edge computing module 103 is further used to:

[0130] S410, when the new layer is rolled for the first time, the loose paving thickness is detected in real time, and when it is detected that the loose paving thickness does not meet the preset value range ΔH, a loose paving alarm message is generated, and the preset value range ΔH is determined by the construction standard;

[0131] S420, when the current rolling construction of the new layer or the old layer is completed, a compaction test report is generated in combination with historical data;

[0132] S430. Based on preset conditions, when the compaction detection report does not meet preset requirements, a compaction alarm message is generated.

[0133] Exemplarily, the edge computing module 103 monitors the loose thickness of the new layer during the initial rolling in real time through the elevation analysis unit. When the loose thickness detection value exceeds the preset range of the construction standard, such as ±5% of the design thickness, an alarm is triggered. At the end of the construction, a sliding window mean filtering algorithm is used to fuse the current rolling trajectory data with the historical layer thickness information to generate a test report containing the compaction compliance rate, thickness uniformity index and interlayer bonding degree. When the key indicators in the compaction test report do not meet the preset requirements, the compaction alarm information can be pushed to the terminal device through the MQTT (message queue telemetry transmission) protocol. Through real-time monitoring and alarm mechanisms, it is ensured that the loose thickness and compaction effect during the construction process meet the preset standards, thereby improving the construction quality and efficiency.

[0134] In one embodiment, the working method of the intelligent display module 104 includes:

[0135] S510, obtaining the thermal map color mapping rules corresponding to the loose paving thickness, the number of compaction passes, and the compaction thickness index;

[0136] S520, based on the color mapping rule of the thermal map, and according to the positioning data and the compaction trajectory, the number of compaction passes and the loose paving thickness information, image rendering is performed to generate a thermal trajectory map of the loose paving thickness, the number of compaction passes and the compaction thickness;

[0137] S530, using a thermal map overlay technology to integrate the thermal trajectory maps of compaction passes, loose paving thickness, and compaction thickness to generate a thermal trajectory map of compaction quality information;

[0138] S540, using SLAM technology, dynamically update the image on the construction display device based on the compaction quality information thermal trajectory map.

[0139] For example, the intelligent display module 104 can mark the loose paving thickness deviation, compaction pass difference and compaction thickness error with different colors based on the HSV color space coding rule, such as green for qualified, yellow for slight deviation, and red for serious over-limit, to form a heat map. A heat map is a visualization tool that uses color to express data density, and is mainly used to display the distribution and concentration of data. These three types of thermal trajectory maps are integrated into a comprehensive quality distribution map through graphic fusion technology. Through SLAM (Simultaneous Localization and Mapping) technology, combined with the real-time positioning information of the roller, the construction scene map is constructed. The quality status map of the construction area is dynamically updated on the display screen in the cab based on the real-time positioning information and the construction scene map. For example, the multi-dimensional heat map is weighted superimposed using Alpha blending technology (an image processing technology, mainly used to achieve image transparency and blending effects), and the ORB-SLAM2 algorithm is used to build a sparse point cloud map of the construction scene. The AR heat map is aligned with the real scene through EPnP (Efficient Perspective-n-Point, a method for solving PnP problems) pose estimation, and the three-dimensional compaction quality distribution is dynamically projected on the display screen in the cockpit at a refresh rate of 30fps. Real-time, intuitive display and dynamic update of compaction quality information are achieved.

[0140] In one embodiment, the cloud collaboration platform 105 is also used for:

[0141] Obtain the operating data of multiple rollers and generate a full-section roadbed compaction 3D surface model based on the operating data;

[0142] Iteratively optimize the feature parameter library based on federated machine learning technology, which contains data features and model training parameters;

[0143] The hash value of the compaction quality data is written into the Ethereum smart contract to generate an electronic construction log.

[0144] Specifically, the cloud collaboration platform 105 can integrate the GNSS trajectory data and vibration energy distribution of multiple devices, and use the Delaunay triangulation algorithm to construct a three-dimensional surface model that reflects the overall compaction state of the roadbed. At the same time, the federated machine learning technology (a distributed machine learning method that exchanges model parameters rather than original data in an encrypted state to protect data privacy) is used to jointly optimize core algorithm models such as thickness prediction and trajectory analysis to improve detection accuracy. For example, a horizontal federation framework is used to upload model gradients, add Gaussian noise, aggregate in the cloud, and optimize the global model once a month. After encrypting key quality data, an electronic file with a timestamp is generated, and through blockchain technology, a digital construction record that cannot be tampered with and can be permanently traced is formed. Accurate monitoring, quality traceability and data security of roadbed compaction construction are achieved.

[0145] In one embodiment, a self-learning optimization module is further included, which is used to:

[0146] S610, obtaining matching data sets of vibration signals and manual detection results under different working conditions to form a knowledge base;

[0147] S620, training a machine learning model using a knowledge base;

[0148] S630, dynamically adjusting the feature extraction threshold of the vibration signal and the weight parameter of the learning model using a Bayesian optimization algorithm;

[0149] S640, based on the field data, using the learning model to generate prediction information and calculate system confidence;

[0150] S650: When the system confidence is less than a preset value, trigger the on-site calibration process and update the knowledge base.

[0151] Specifically, the system achieves intelligent calibration by integrating machine learning and dynamic optimization: collecting vibration signals and manual inspection results under different working conditions, such as collecting roller inspection data and manual inspection data using the sand filling method and ring knife method, building a training database, using the random forest algorithm to train the prediction model, using Bayesian optimization technology, with the goal of minimizing the model prediction error, and automatically adjusting the feature extraction threshold and model weight parameters of the vibration signal. During on-site construction, the confidence of the model output is calculated in real time. If the probability value is less than 85%, the equipment is triggered to automatically shut down and prompt manual re-inspection. At the same time, the re-inspection data is fed back to the knowledge base to iteratively update the model, forming a closed-loop learning mechanism of "data accumulation-model optimization-abnormal self-inspection" to achieve continuous optimization of the model and high-precision prediction.

[0152] In one embodiment, the detection report includes the warning number, event level, warning type, warning message and timestamp:

[0153] The warning pile number is used to identify the specific location where the warning occurs;

[0154] The event level is used to classify warning events into three levels: severe, moderate, and minor according to the severity of the warning;

[0155] Warning type is used to specify the type of warning, including unqualified compaction passes and unqualified compaction thickness;

[0156] Warning messages are used to describe the specific content of the warning, including the cause of the warning and possible impact.

[0157] Specifically, the inspection report contains five types of core information: Warning pile number, which can be used to accurately locate the specific construction location where the warning occurred. Event level, which divides warnings into three levels: severe, moderate, and minor according to severity, to guide differentiated response priorities. Warning type, which can clarify the problem category, including two types of engineering indicator abnormalities: unqualified compaction times and unqualified compaction thickness. Warning message, which describes the cause of the warning in detail, such as insufficient rolling times and potential impacts, such as the risk of roadbed settlement. Timestamp, which records the warning trigger time and supports construction quality traceability and process review. Through structured report content, construction problems can be quickly located, graded and handled, and closed-loop managed to improve quality control efficiency.

[0158] The roadbed compaction detection system provided by this application has overcome the industry problems of low efficiency, poor accuracy and unreliable data of traditional manual detection through multi-module collaborative innovation. The centimeter-level GNSS high-precision positioning module and the data acquisition module 102 realize the perception of all elements of the construction process, and the edge computing module 103 is used to generate compaction trajectories, pass statistics and thickness indicators in real time, saving the time of manual detection. The intelligent display module 104 dynamically maps the quality deviation through the heat map, and the compaction status of the entire roadbed is clear at a glance. The cloud platform builds tamper-proof construction archives based on blockchain, realizes multi-device data integration and federated learning optimization, reduces the model prediction error, the elevation analysis unit stratifies the construction layer, and the experimental calibration module optimizes the model parameters through dual-zone comparison, which can adapt to various working conditions. The self-learning module learns according to the updated data and keeps pace with the times. A closed-loop system of perception-analysis-feedback-traceability is formed, which ultimately improves the compaction qualification rate and reduces the comprehensive construction cost, providing a full-chain solution for intelligent road construction.

[0159] In order to further illustrate the solution of the embodiment of the present application, a specific example is provided below.

[0160] A roadbed compaction detection system includes a high-precision positioning module, a data acquisition module 102, an edge computing module 103, an intelligent display module 104 and a cloud collaboration platform 105:

[0161] High-precision positioning module, used to obtain centimeter-level positioning data of the roller in real time;

[0162] Specifically, the high-precision positioning module establishes a Beidou satellite CORS station to perform real-time differential calculation of the location information within the service range, achieve centimeter-level positioning of the roller's rolling trajectory, and provide a spatial reference for trajectory generation.

[0163] The data acquisition module 102 is used to collect vertical vibration signals generated by the roller during the roadbed compaction process;

[0164] Specifically, the data acquisition module 102 includes a vibration sensor for collecting acceleration information of the vibration wheel and generating a vibration signal.

[0165] The edge computing module 103 is used to generate quality data of compaction trajectory, compaction passes, loose paving thickness and compaction thickness in real time according to the positioning data and vibration signals;

[0166] The edge computing module 103 includes a data processing unit, which is used to process the split data in real time according to a specific algorithm to obtain quality data of compaction trajectory, compaction passes, loose laying thickness and compaction thickness.

[0167] Intelligent display module 104, used to map compaction quality data to a construction display device through a thermal map;

[0168] Specifically, the intelligent display module 104 includes an intelligent display terminal, which generates a thermal map based on the mapping rules of the quality data of the compaction trajectory, the number of passes and the thickness, and displays the thermal map on the intelligent display terminal. For example, the following mapping rules are used: red represents a compaction thickness of less than 15 cm or a loose thickness of less than 18 cm, yellow represents a compaction thickness of 15-25 cm or a loose thickness of 18-30 cm, and green represents a compaction thickness of more than 25 cm or a loose thickness of more than 30 cm. The HSV color space encoding rule is used, and the linear difference from red to green is equally divided into 8 color levels, and each color level corresponds to 1 to 8 rolling passes, and the generated thermal map is mapped to the construction display device.

[0169] In this example, the high-precision positioning module is used to obtain centimeter-level positioning data. The data acquisition module 102 collects vibration signals. The edge computing module 103 generates compaction quality data in real time. The intelligent display module 104 intuitively displays these quality data through airborne equipment, thus achieving accurate monitoring and quality control of the entire roadbed compaction process.

[0170] In this example, the roadbed compaction detection system can be powered by the roller and started synchronously with the roller. After starting, the equipment runs automatically. The software startup and network connection do not require manual operation. The user only needs to view the compaction quality indicators through the intelligent display terminal, which reduces the user's operating burden.

[0171] This example also includes an experiment adjustment module, and the experiment scheme adopted by the experiment adjustment module is:

[0172] The experimental area was divided into qualified area and unqualified area along the center line. The length of both areas was 50 meters and the width was 5 meters. The loose paving thickness of the qualified area was 20 cm and that of the unqualified area was 40 cm.

[0173] Set the number of rolling times for qualified areas to 4 and the number of rolling times for unqualified areas to 2;

[0174] Obtain quality data of compaction passes, loose paving thickness and compaction thickness in roller experiments;

[0175] Real-time alarm feedback is provided based on the quality data and qualified standards of compaction passes, loose laying thickness and compaction thickness.

[0176] Specifically, by dividing the experimental area into qualified areas and unqualified areas, and setting different loose paving thicknesses and rolling times, the experimental adjustment module can simulate different compaction conditions in actual construction to ensure the representativeness and comprehensiveness of the experimental data. The experimental adjustment module can obtain the quality data of compaction times, loose paving thickness and compaction thickness in real time, and compare them with the qualified standards to promptly discover quality problems in the compaction process. In a roadbed pressure test detection system in this example, the qualified standard for loose paving thickness is 18-30cm, the qualified standard for compaction thickness is 15-25cm, and the number of rolling times is greater than 4-8 times. When the detected quality data does not meet the qualified standards, the system can issue an alarm in real time. The specific alarm mode is: when the roller enters the new layer to work, when the roller is rolling in the unqualified loose paving area, the loose paving thickness is alarmed in real time according to the pile number. When the roller rolls in the qualified loose paving area, the loose paving thickness does not alarm. When the roller stops or completes this work, the compaction times and compaction thickness are alarmed. The alarm can remind construction personnel to adjust construction parameters in time to avoid further expansion of quality problems and ensure construction quality.

[0177] The edge computing module 103 of this example further includes an elevation analysis unit, which is used to:

[0178] Based on the real-time positioning data of adjacent working areas, determine the relationship between layers:

[0179] If the difference in elevation between the current working interval and the adjacent working interval is ≥12cm, the working interval with a higher elevation is defined as the new layer, and the working interval with a lower elevation is defined as the old layer;

[0180] If the elevation difference between the current working area and the adjacent working area is less than 12 cm, they are considered to be on the same floor;

[0181] Generate inter-layer inspection reports based on working intervals and inter-layer relationships.

[0182] For example, if the elevation difference between a single small area and the adjacent area data exceeds 12 cm, it can be considered that the point has entered a new layer. A roadbed compaction detection system in this example can effectively detect the relationship between the new layer or the old layer of the roadbed, thereby improving construction efficiency.

[0183] In this example, a roadbed compaction detection system uses the detected plane coordinates, roadbed elevation, mechanical vibration status and other construction big data, and through real-time data transmission and processing, it realizes real-time monitoring and control of key indicators such as roadbed compaction times, compaction thickness, and compaction degree. It solves the problems of low efficiency, poor representativeness, and delayed detection of traditional detection methods such as sand filling method and ring knife method, and has the advantages of fast real-time, full coverage, strong representativeness, and visual results.

[0184] It should be understood that, although the various steps in the flowcharts involved in the above-mentioned embodiments are displayed in sequence according to the indication of the arrows, these steps are not necessarily executed in sequence according to the order indicated by the arrows. Unless there is a clear explanation in this article, the execution of these steps does not have a strict order restriction, and these steps can be executed in other orders. Moreover, at least a part of the steps in the flowcharts involved in the above-mentioned embodiments can include multiple steps or multiple stages, and these steps or stages are not necessarily executed at the same time, but can be executed at different times, and the execution order of these steps or stages is not necessarily to be carried out in sequence, but can be executed in turn or alternately with other steps or at least a part of the steps or stages in other steps.

[0185] The above-mentioned embodiments only express several implementation methods of the embodiments of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the patent of the embodiments of the present application. It should be pointed out that, for ordinary technicians in this field, several variations and improvements can be made without departing from the concept of the embodiments of the present application, and these all belong to the protection scope of the embodiments of the present application.

Claims

1. A roadbed compaction detection system, characterized in that: Including high-precision positioning module, data acquisition module, edge computing module, intelligent display module and cloud collaboration platform: The high-precision positioning module is used to obtain centimeter-level positioning data of the roller in real time; The data acquisition module is used to collect vertical vibration signals generated by the roller during the roadbed compaction process; The edge computing module is used to generate quality data of compaction trajectory, compaction passes, loose paving thickness and compaction thickness in real time according to the positioning data and the vibration signal; The intelligent display module is used to map the compaction quality data to the construction display device through a thermal map; The cloud-based collaborative platform is used to integrate the collaborative operation data of multiple devices and realize quality traceability based on blockchain technology.

2. A roadbed compaction detection system according to claim 1, characterized in that: The edge computing module includes a trajectory generation unit, a pass calculation unit and a thickness modeling unit: The trajectory generating unit is used to generate a compaction trajectory by fitting the positioning data using a B-spline curve; The pass calculation unit is used to generate the compaction pass number by dividing the working interval into preset sizes and counting the number of times the compaction track covers each interval; The thickness modeling unit is used to establish a relationship model between loose paving thickness and compacted thickness based on vibration signal characteristic parameters and experimental data.

3. A roadbed compaction detection system according to claim 2, characterized in that: The compacted thickness is calculated by the following method: Use the following formula to calculate the autocorrelation function R(k) corresponding to the vibration response signal: Where x[n] represents the vibration response signal, N represents the signal length, and k represents the lag time; Use the following formula to identify the fundamental frequency f0 of the vibration response signal x[n]: Among them, k p represents the hysteresis of the first peak of R(k); Calculate the Fourier transform X(f) of the vibration response signal x[n] using the following formula: Among them, e represents a natural number, f represents a frequency, and j represents an imaginary unit; Identify the fundamental frequency f0 and its harmonics f n =nf0, where n=1, 2, ...; Use the following formula to calculate the amplitude A of the identified fundamental frequency and harmonics 基 and the amplitude of the nth harmonic A 谐n : A 基 =|X(f0)|,A 谐n =|X(nf0)|; Calculate the compaction thickness H using the following formula: Among them, v s represents the roller travel speed (m / s), which is obtained by differential calculation of GNSS positioning data, q represents the quality index of the roadbed, and a, b and c represent parameters determined by experimental calibration.

4. A roadbed compaction detection system according to claim 3, characterized in that: The invention also includes an experiment adjustment module, wherein the experiment adjustment module is used to: The experimental area is divided into qualified area and unqualified area along the center line. The width of the two areas is 5 meters. The loose paving thickness x of the qualified area is based on the maximum particle size D of the material. m and design target compaction H d Sure: x=0.4·D m +0.2·H d Among them, the value range of x is 15cm to 35cm; The loose paving thickness of the unqualified area is y=x+Δh, where Δh is determined by the construction standard; Set the number of rolling times w in the qualified area, w is based on the target compaction degree K t calculate: Where, μ represents the single rolling efficiency coefficient, μ = 0.15, K t ≥95%; The number of times the unqualified area is rolled is z=w-Δj, where Δj is the preset number of deviations; Calculating the quality data deviation of the compaction track, compaction passes, loose paving thickness and compaction thickness detected by the roller and manually, and obtaining detection deviation information; According to the detection deviation information, the calculation parameters of the edge computing module are adjusted using the least squares method.

5. A roadbed compaction detection system according to claim 2, characterized in that: The edge computing module further includes an elevation analysis unit, which is used to: Based on the real-time positioning data of the adjacent working intervals, the relationship between layers is determined: If the difference in elevation between the current working interval and the adjacent working interval is ≥12cm, the working interval with a higher elevation is defined as a new layer, and the working interval with a lower elevation is defined as an old layer; If the elevation difference between the current working area and the adjacent working area is less than 12 cm, they are considered to be on the same floor; An inter-layer detection report is generated according to the working interval and the inter-layer relationship.

6. A roadbed compaction detection system according to claim 5, characterized in that: The edge computing module is also used for: When the new layer is rolled for the first time, the loose paving thickness is detected in real time, and when it is detected that the loose paving thickness does not meet the preset value range ΔH, a loose paving alarm message is generated, and the preset value range ΔH is determined by the construction standard; When the current rolling construction of the new layer or the old layer is completed, a compaction test report is generated in combination with historical data; Based on preset conditions, when the compaction detection report does not meet preset requirements, compaction alarm information is generated.

7. A roadbed compaction detection system according to claim 1, characterized in that: The working method of the intelligent display module includes: Get the thermal map color mapping rules corresponding to the loose paving thickness, compaction passes, and compaction thickness indicators; Based on the color mapping rules of the thermal map, and according to the positioning data and the compaction trajectory, the number of compaction passes and the loose paving thickness information, image rendering is performed to generate a thermal trajectory map of the loose paving thickness, the number of compaction passes and the compaction thickness; The thermal trajectory diagrams of compaction passes, loose paving thickness and compaction thickness are integrated by using thermal map superposition technology to generate a thermal trajectory diagram of compaction quality information; By using SLAM technology, the image on the construction display device is dynamically updated according to the thermal trajectory diagram of the compaction quality information.

8. A roadbed compaction detection system according to claim 2, characterized in that: The cloud collaboration platform is also used for: Acquire operation data of multiple rollers, and generate a full-section roadbed compaction three-dimensional surface model according to the operation data; Iteratively optimize a feature parameter library based on federated machine learning technology, wherein the feature parameter library includes data features and model training parameters; The hash value of the compaction quality data is written into the Ethereum smart contract to generate an electronic construction log.

9. A roadbed compaction detection system according to claim 1, characterized in that: Also includes self-learning optimization modules for: Acquire matching data sets of the vibration signals and manual detection results under different working conditions to form a knowledge base; Training a machine learning model using the knowledge base; Dynamically adjusting the feature extraction threshold of the vibration signal and the weight parameters of the learning model using a Bayesian optimization algorithm; Based on the field data, the learning model is used to generate prediction information and calculate the system confidence; When the system confidence is less than a preset value, an on-site calibration process is triggered and the knowledge base is updated.

10. A roadbed compaction detection system according to claim 6, characterized in that: The detection report includes the warning pile number, event level, warning type, warning message and timestamp: The warning stake number is used to identify the specific location where the warning occurs; The event level is used to classify warning events into three levels: severe, moderate and minor according to the severity of the warning; The warning type is used to specify the type of warning, including unqualified compaction passes and unqualified compaction thickness; The warning message is used to describe the specific content of the warning, including the cause of the warning and possible impact.

Images