Fusion air-rail multidimensional information maglev line track digital twin state monitoring method

Abstract

The present application belongs to the technical field of track monitoring, and provides a fusion air-rail-ground multi-dimensional information maglev line track digital twin state monitoring method, which comprises: establishing a three-dimensional model of a segmented line track based on aerial photography information of a UAV and BIM technology; obtaining surface feature data of the selected segmented line track by using a cross-section scanner, and combining attitude and position information of an inertial navigation system to reconstruct the three-dimensional model of the selected segmented line track; finally splicing the reconstructed three-dimensional model and the un-reconstructed segmented line track model to obtain a line track geometric model, improve the modeling accuracy, construct a database based on the qualified holographic state data, combine the completed geometric model to construct a behavior and rule model of the maglev line track digital base, and use a Kalman filter fusion algorithm to fuse the isomorphic and heterogeneous holographic state data, so as to realize diversified intelligent monitoring of the maglev line track.

Description

Technical Field

[0001] This invention belongs to the field of track monitoring technology, specifically a method for monitoring the digital twin status of maglev line tracks by integrating multi-dimensional information from air, track, and ground. Background Technology

[0002] Maglev transportation represents the forefront of rail transit technology development in the post-high-speed rail era, showcasing a nation's advanced manufacturing and R&D capabilities. Maglev transportation is highly integrated and tightly coupled, combining features of civil engineering structures and electromechanical facilities in its track systems. Therefore, intelligent monitoring of maglev transportation is of great significance. From a macro perspective, drone technology provides fundamental safety assurance along maglev lines and is a prerequisite for the integration of digital twins. After drones ensure the safety of the maglev line's track clearance limits, data collection and analysis using on-orbit monitoring equipment address existing maglev line issues and meet high-precision requirements. Compensation modeling from an on-orbit microscopic perspective, along with ground-based fixed-point monitoring, effectively compensates for the insufficient analysis of maglev line structural performance at the aerial and track levels, ensuring overall normal structural performance under all weather conditions from a health monitoring perspective.

[0003] At the data fusion level, multi-source data collected by multiple sensors is preprocessed. Interpolation algorithms and filtering methods (such as mean filtering) are used to address data loss and noise issues. After data cleaning, data fusion algorithms are used to fuse the original multi-source data as the foundation for fusion information. In feature fusion technology, feature extraction algorithms (such as PCA and LDA) are used to reduce the dimensionality and extract features from multi-source data, fusing feature data from different information sources. This transforms data of different dimensions into the same feature space, and deep learning is combined to perform unified modeling of multi-source features, meeting the high-precision state information requirements of the model. At the decision-making level, combining the data analysis results from different dimensions of air, rail, and ground, various decision fusion methods (such as DS evidence theory and voting mechanisms) are used to integrate and output multi-dimensional decisions based on the cleaned data.

[0004] In the current technology, the inspection and maintenance of maglev operating track lines still mainly rely on patrols under the safety beams during operating hours and manual inspections on the beams during maintenance windows. This has shortcomings such as low efficiency, limited maintenance and inspection time, poor working environment, and inability to accurately reflect the system status in real time, which cannot meet the needs of maglev transportation development.

[0005] To address this, the present invention provides a digital twin status monitoring method for maglev line tracks that integrates multi-dimensional information from air, rail, and ground. Considering the high linearity of maglev lines and the potential safety hazards associated with traversing urban areas, the method employs unmanned aerial vehicles (UAVs) for monitoring and image acquisition. Furthermore, to address the high precision requirements of the tracks, on-orbit equipment such as structured light scanning and cross-section scanners are used to provide detailed monitoring of the facilities. This is combined with ground-based health monitoring (industrial cameras, ground-based fixed-point monitoring) to compensate for the limitations of on-orbit equipment in monitoring structural performance and on-orbit inspections during track windows, effectively improving the monitoring efficiency and level of maglev line tracks.

[0006] This monitoring method first involves multi-dimensional data collection and processing from both the air and ground levels, analyzing data from multiple angles and in all dimensions to monitor track status from overall to detailed aspects, achieving comprehensive status monitoring and early warning. Then, a progressive information fusion process ensures the orderly connection between the refined processing of basic data, feature extraction, and final comprehensive decision-making, guaranteeing the scientific validity and reliability of the decisions. Finally, a digital twin platform serves as the hub for information transmission and processing, organically combining the fusion and analysis results of multi-dimensional data with the final monitoring decisions, ensuring a closed-loop and efficient monitoring process, as well as a comprehensive and intuitive monitoring interface for users. Summary of the Invention

[0007] In order to overcome the shortcomings of the prior art, at least one technical problem raised in the background art is solved.

[0008] The technical solution adopted by this invention to solve its technical problem is: a method for monitoring the digital twin status of maglev line tracks by integrating multi-dimensional information from air, rail, and ground, comprising:

[0009] Based on aerial photography data from UAVs and BIM technology, 3D models of segmented track lines are established and initially stitched together. The overlapping areas and their representational data between adjacent segmented track 3D models are obtained, including the area of ​​the overlapping area and the number of point clouds. Based on the processing and analysis of this representational data, segmented track lines are selected. Surface feature data of the selected segmented track lines are acquired using a cross-section scanner. Combined with attitude and position information from an inertial navigation system, the 3D models of the selected segmented track lines are reconstructed using BIM technology. Finally, the high-precision 3D models of the reconstructed segmented track lines are stitched together with the unreconstructed models and exported in a universal format to obtain the geometric model of the track lines.

[0010] Based on the multi-dimensional information acquisition system of air-rail-ground, holographic state data of maglev line track is obtained. Based on the processing and analysis of holographic state data, missing data values ​​are obtained. Based on the missing data values, the holographic state data is evaluated. Based on the qualified holographic state data, a database is constructed. Combined with the completed geometric model, the behavior and rule model of the digital base of maglev line track is constructed using BIM technology.

[0011] Based on qualified holographic state data, the Kalman filter fusion algorithm is used to fuse information between isomorphic and heterogeneous holographic state data within the holographic state data.

[0012] A smart monitoring platform for maglev lines is constructed by integrating qualified holographic status data with geometric, behavioral, and rule models of the track.

[0013] As a further technical solution of the present invention, the process of selecting the segmented track is as follows:

[0014] The obtained point cloud distribution representation value DYB and overlap degree CD are processed using the formula: The model reconstruction value CG is obtained, where a1 and a2 are preset scaling coefficients;

[0015] Compare the model reconstruction value CG with the model reconstruction threshold;

[0016] If the model reconstruction value CG is greater than or equal to the model reconstruction threshold, it indicates that the overlap of the overlapping area is high and the point cloud distribution is uniform and the density is qualified, so no operation is performed.

[0017] If the model reconstruction value CG is less than the model reconstruction threshold, it indicates that the overlap of the overlapping area is low and the point cloud distribution is uneven and the density is unqualified. In this case, the two segmented track sections corresponding to the overlapping area are selected for model reconstruction.

[0018] A further technical solution of the present invention is as follows: the method for obtaining the overlap degree CD is:

[0019] Key feature points are extracted from the model using feature point detection and description algorithms, and similar feature points in adjacent models are found using feature point matching algorithms. Overlapping regions between models are obtained through feature point indicators.

[0020] Obtain the area of ​​the overlapping region and the area of ​​the minimum bounding rectangle corresponding to the model. Calculate the ratio between the area of ​​the overlapping region and the area of ​​the minimum bounding rectangle corresponding to the model to obtain the overlap degree, and mark it as CD.

[0021] The minimum bounding rectangle refers to the smallest rectangle that can enclose all target points, whose width and height are determined by the maximum and minimum values ​​of the target points.

[0022] A further technical solution of the present invention is as follows: the point cloud distribution representation value DYB is obtained in the following way:

[0023] The ratio of prominent overlap (CS) and the value of prominent overlap (CZ) are processed using the following formula: The non-uniform point cloud density value DJ is obtained, where s1 and s2 are both preset scaling coefficients;

[0024] The point cloud distribution performance value is obtained by summing the point cloud density non-uniformity value DJ with the point cloud deviation ratio, and then labeled as DYB.

[0025] The point cloud deviation ratio is obtained as follows:

[0026] Point cloud analysis is performed on the overlapping area using point cloud processing software to obtain the number of point clouds in the overlapping area. The point cloud density of the overlapping area is obtained by comparing the number of point clouds in the overlapping area with the area of ​​the overlapping area. The difference between the point cloud density of the overlapping area and the set point cloud density of the overlapping area is calculated, and the absolute value of the difference is taken to obtain the point cloud density difference of the overlapping area. The point cloud density difference of the overlapping area is then compared with the set point cloud density of the overlapping area to obtain the point cloud deviation ratio.

[0027] A further technical solution of the present invention is as follows: the method for obtaining the prominent overlap ratio CS is:

[0028] The overlapping region is divided into several overlapping sub-regions with equal areas. The number of point clouds in each overlapping sub-region is obtained and then compared with the area of ​​the overlapping sub-region to obtain the point cloud density of the overlapping sub-region.

[0029] The point cloud density of all overlapping sub-regions is summed and averaged to obtain the average point cloud density. The point cloud density of each overlapping sub-region is then compared with the average point cloud density, and the absolute value of the difference is taken to obtain the point cloud density deviation of the overlapping sub-region. The point cloud density deviation of all overlapping sub-regions is summed and averaged to obtain the average point cloud density deviation. The point cloud density deviation of the overlapping sub-region is then compared with the average point cloud density deviation.

[0030] If the point cloud density deviation of an overlapping sub-region is greater than the mean point cloud density deviation, then the overlapping sub-region is marked as a prominent overlapping sub-region.

[0031] If the point cloud density deviation of the overlapping sub-region is less than or equal to the mean point cloud density deviation, then the overlapping sub-region is marked as a non-prominent overlapping sub-region.

[0032] The number of highlighted overlapping sub-regions is counted, and the ratio of this number to the total number of overlapping sub-regions is calculated to obtain the highlighted overlap ratio, which is then labeled as CS.

[0033] A further technical solution of the present invention is as follows: the method for obtaining the degree of overlap CZ is as follows:

[0034] The mean deviation of the point cloud density between the overlapping sub-regions and the mean point cloud density is processed to obtain the mean deviation of the point cloud density of the overlapping sub-regions. The mean deviation of the point cloud density of all overlapping sub-regions is summed and averaged to obtain the relative deviation of the point cloud density. The ratio of the relative deviation of the point cloud density to the mean point cloud density deviation is processed to obtain the degree of overlap, and it is labeled as CZ.

[0035] As a further technical solution of the present invention, the process of acquiring the holographic state data includes:

[0036] A1. Technical means of high-altitude measurement: UAVs conduct regular automatic flight inspections to achieve rapid detection of the status information of cable trenches, switch houses, and substation line auxiliary facilities;

[0037] A2, Technical means of dynamic track inspection: on-orbit machine vision inspection and structured light scanning. The on-orbit machine vision inspection system refers to the use of industrial cameras and edge computers as the main components, based on high-speed shooting and machine learning algorithms to identify the following information: structural seam recognition, foreign object intrusion in functional areas, stator surface wear, and substructure defects, and record the mileage information and anomaly type of the identified location; the structured light profile scanner is used to accurately scan the stator surface, and the stator surface deviation and fold angle are measured by point cloud processing to describe the geometric shape and position deviation.

[0038] A3. Select typical track beams to establish a fixed structural health monitoring system and build an all-weather fixed-point monitoring network for track structure and environmental conditions. The monitoring objects include: track beam end inclination angle, mid-span vibration response characteristics, surface temperature distribution, and environmental conditions.

[0039] A further technical solution of the present invention is as follows: the method for obtaining the missing data value QP is:

[0040] Set the acquisition period, divide the acquisition period into several time analysis nodes, acquire the holographic state data at the time analysis nodes, and compare the holographic state data at the time analysis nodes with the standard holographic state data.

[0041] If the total amount of data contained in the holographic state data at the time analysis node is not equal to the total amount of data contained in the standard holographic state data, then the time analysis node will be marked as an abnormal acquisition node.

[0042] If the total amount of data contained in the holographic state data at the time analysis node is equal to the total amount of data contained in the standard holographic state data, then no processing is performed;

[0043] The number of abnormal data collection nodes is counted, and the ratio of this number to the number of time analysis nodes is calculated to obtain the abnormal data collection ratio, which is then labeled as SL.

[0044] The difference between the total amount of data contained in the holographic state data at the abnormal acquisition node and the total amount of data contained in the standard holographic state data is processed to obtain the data missing amount at the abnormal acquisition node. The data missing amounts at all abnormal acquisition nodes are summed and averaged to obtain the average data missing amount. The ratio of the average data missing amount to the total amount of data contained in the standard holographic state data is processed to obtain the abnormal acquisition missing ratio, which is marked as QS.

[0045] The obtained outlier count ratio (SL) and outlier missing ratio (QS) are processed using the following formula: The missing data value QP is obtained, where z1 and z2 are preset scaling coefficients.

[0046] As a further technical solution of the present invention, the process of evaluating the holographic state data according to the missing data value is as follows: comparing the missing data value QP with the missing data threshold.

[0047] If the missing data value QP is greater than or equal to the missing data threshold, it indicates that the holographic state data collected during the acquisition period is highly missing and is unqualified.

[0048] If the missing data value QP is less than the missing data threshold, it indicates that the degree of missing data in the holographic state collected during the acquisition period is low and acceptable.

[0049] As a further technical solution of the present invention, the specific process of integrating and constructing the intelligent monitoring platform for maglev line tracks is as follows:

[0050] H1, using MATLAB or Python data analysis software to analyze the holographic state data of the maglev line track system, specifically as follows:

[0051] Track geometry: including track dimensions, alignment, levelness, and elevation difference;

[0052] Structural health status: Monitor the stress, strain, vibration, and temperature parameters of the track beams, supports, and connecting structural components;

[0053] Environmental parameters: such as temperature, humidity, wind speed, and rainfall;

[0054] Equipment status: Operating status and performance parameters of signal system, power supply system, and communication system equipment;

[0055] H2 utilizes decentralized edge computing technology to perform preliminary data preprocessing at the device level, thereby reducing data transmission volume, improving the efficiency of information transmission between the cloud and the device, and reducing the processing pressure on the cloud. Specifically:

[0056] Gross error assessment: At the data acquisition point, preliminary screening is performed to remove obviously erroneous data;

[0057] Data cleaning: removing duplicate data, filling in missing values, handling outliers, and improving data quality;

[0058] H3 uses a cloud computing platform as its foundation, imports the railway track model, and completes the construction of a smart monitoring platform, specifically including:

[0059] Model demonstration: Import the geometric model, behavior and rule model of the track into the cloud platform to demonstrate the structure and status of the maglev track system;

[0060] Data Download: Provides a data download interface, allowing users to download raw or processed data as needed;

[0061] Database Management: Design and implement an efficient database management system for storing and managing various types of data for the maglev track system;

[0062] Login and Access Control: Provides user login and access control functions to ensure system security and data confidentiality.

[0063] The beneficial effects of this invention are as follows:

[0064] 1. Based on aerial photography data from UAVs and BIM technology, a 3D model of the segmented track is established and initially stitched together. The overlapping areas and their representation data between adjacent segmented track 3D models are obtained. Based on the processing and analysis of the representation data, segmented track models are selected. Surface feature data of the selected segmented track is acquired using a cross-section scanner. Combined with attitude and position information from an inertial navigation system, a high-precision 3D model of the selected segmented track is reconstructed using BIM technology. The reconstructed high-precision 3D model of the segmented track is then finally stitched together with the unreconstructed segmented track model and exported to a universal format to obtain the track geometry model. UAV aerial photography modeling provides a good foundation for macroscopic modeling, enabling overall modeling of the maglev line and surrounding facilities, and achieving relevant protection within building clearances. Due to the high mobility of UAVs, the modeling process has high modeling efficiency and protection capabilities. Furthermore, the combination of a cross-section scanner and an inertial navigation system enables the compensated reconstruction of the track model, meeting the precision requirements for fine-grained maglev modeling. This modeling process for maglev tracks has practical engineering significance and is conducive to intelligent monitoring of the track.

[0065] 2. Based on a multi-dimensional information acquisition system encompassing air, rail, and ground, holographic status data of maglev line tracks is obtained. This data is then processed and analyzed to identify missing values. Based on these missing values, the holographic status data is evaluated. A database is constructed using the evaluated holographic status data. Combined with the completed geometric model, BIM technology is used to construct a behavioral and rule model of the digital base of the maglev line track. Based on the qualified holographic status data, a Kalman filter fusion algorithm is used to fuse isomorphic and heterogeneous holographic status data. This invention achieves the construction of a behavioral and rule model of the digital base of the maglev line track, enabling the monitoring of the maglev line track and the fusion of diverse holographic status data, thus realizing diversified intelligent monitoring of the maglev line track.

[0066] 3. Based on the collected qualified holographic status data and the geometric, behavioral, and rule models of the track, a smart monitoring platform for maglev lines is constructed. First, image acquisition by drones enables precise construction of the track model, improving model building efficiency. Second, drone-based data acquisition overcomes the problems of untimely and inefficient monitoring caused by long track lengths. Furthermore, data acquisition through structured light scanning and cross-section scanners for model building and fusion monitoring meets the high precision requirements of track monitoring. The health monitoring system, composed of industrial cameras and ground-based fixed-point monitoring, overcomes the limitations of traditional track structure self-monitoring. This invention achieves intelligent monitoring of the track from multiple dimensions, improving the convenience and efficiency of maglev track monitoring. First, information is collected and processed from both air, track, and ground dimensions to construct the track model while simultaneously achieving detailed and comprehensive monitoring of the track status. Then, data processing ensures the orderly connection of basic data, guaranteeing accuracy and scientific validity, such as data anomaly analysis. Finally, a digital twin platform organically combines the fusion and analysis results of multi-dimensional data with the final monitoring decisions, forming a complete smart monitoring system. Attached Figure Description

[0067] The invention will now be further described with reference to the accompanying drawings.

[0068] Figure 1 This is a flowchart illustrating the steps of the method for monitoring the digital twin status of maglev line tracks by integrating multi-dimensional information from air, rail, and ground, as described in this embodiment of the invention.

[0069] Figure 2 This is a flowchart of the judgment process for three-dimensional model reconstruction in the maglev line track digital twin status monitoring method that integrates multi-dimensional information of air, rail, and ground, as described in this embodiment of the invention.

[0070] Figure 3This is a model comparison diagram after modeling is completed in an embodiment of the present invention;

[0071] Figure 4 This is a schematic diagram of the structured light scanning monitoring used in an embodiment of the present invention;

[0072] Figure 5 This is a schematic diagram illustrating the use of an industrial camera for detection in an embodiment of the present invention;

[0073] Figure 6 This is a schematic diagram illustrating the use of ground-based fixed-point monitoring in an embodiment of the present invention;

[0074] Figure 7 This is a block diagram of the maglev line track digital twin status monitoring system that integrates multi-dimensional information from air, rail, and ground, as described in an embodiment of the present invention. Detailed Implementation

[0075] To make the technical means, creative features, objectives and effects of this invention easier to understand, the invention will be further described below in conjunction with specific embodiments.

[0076] Example 1

[0077] like Figure 1 As shown in the embodiment of the present invention, the method for monitoring the digital twin status of maglev line tracks by integrating multi-dimensional information on air, rail, and ground includes:

[0078] Step 1: Based on aerial photography information from UAVs and BIM technology, establish 3D models of segmented track lines and perform preliminary stitching. Obtain the overlapping areas and representation data of adjacent segmented track 3D models after stitching. The representation data includes the area of ​​the overlapping area and the number of point cloud points. Based on the processing and analysis of the representation data, select the segmented track lines. Use a cross-section scanner to obtain the surface feature data of the selected segmented track lines. Combine the attitude and position information of the inertial navigation system and use BIM technology to reconstruct the 3D models of the selected segmented track lines. Finally, stitch the reconstructed high-precision 3D model of the segmented track line with the unreconstructed segmented track model and export it to a general format to obtain the track geometry model.

[0079] It should be noted that the specific operations before constructing the 3D model of the segmented track also include:

[0080] S1, survey the entire mileage of the maglev line, record the areas where drone flight is restricted and the corresponding mileage information, and set up ground control points and make corresponding markings based on this information;

[0081] S2 combines the effective operating time of a single UAV flight, the flight-limited mileage location, and the minimum coverage requirements of the ground control points. Based on the coordinate information in the route design stage, it rationally plans the detection area and the number of flights, and unifies the standard for single flight route design.

[0082] S3 uses GNSS base stations to measure the coordinate information of each control point along the line, which is used for adjustment and correction before the establishment of a three-dimensional model of the segmented track.

[0083] like Figure 2 As shown, in some embodiments, key feature points are extracted from the model using feature point detection and description algorithms (such as SIFT, SURF, ORB, etc.), and similar feature points in adjacent models are found using feature point matching algorithms (such as FLANN, BFMatcher, etc.); overlapping areas between models are obtained through feature point indications.

[0084] Obtain the area of ​​the overlapping region and the area of ​​the minimum bounding rectangle (MBR) corresponding to the model. Calculate the ratio between the area of ​​the overlapping region and the area of ​​the minimum bounding rectangle corresponding to the model to obtain the overlap degree, and mark it as CD.

[0085] The minimum bounding rectangle (MBR) is defined as the smallest rectangle that can enclose all target points (in this scenario, the target points are all vertices of the model), whose width and height are determined by the maximum and minimum values ​​of the target points.

[0086] Point cloud processing software is used to perform point cloud analysis on the overlapping area to obtain the number of point cloud points in the overlapping area. The ratio of the number of point cloud points in the overlapping area to the area of ​​the overlapping area is processed to obtain the point cloud density of the overlapping area. The difference between the point cloud density of the overlapping area and the set point cloud density of the overlapping area is processed, and the absolute value of the difference is taken to obtain the point cloud density difference of the overlapping area. The point cloud density difference of the overlapping area is compared with the set point cloud density of the overlapping area to obtain the point cloud deviation ratio.

[0087] It should be noted that the point cloud density is set in advance by those skilled in the art based on the model building requirements;

[0088] The overlapping region is divided into several overlapping sub-regions with equal areas. The number of point clouds in each overlapping sub-region is obtained and then compared with the area of ​​the overlapping sub-region to obtain the point cloud density of the overlapping sub-region.

[0089] The point cloud density of all overlapping sub-regions is summed and averaged to obtain the average point cloud density. The point cloud density of each overlapping sub-region is then compared with the average point cloud density, and the absolute value of the difference is taken to obtain the point cloud density deviation of the overlapping sub-region. The point cloud density deviation of all overlapping sub-regions is summed and averaged to obtain the average point cloud density deviation. The point cloud density deviation of the overlapping sub-region is then compared with the average point cloud density deviation.

[0090] If the point cloud density deviation of an overlapping sub-region is greater than the mean point cloud density deviation, then the overlapping sub-region is marked as a prominent overlapping sub-region.

[0091] If the point cloud density deviation of the overlapping sub-region is less than or equal to the mean point cloud density deviation, then the overlapping sub-region is marked as a non-prominent overlapping sub-region.

[0092] The number of highlighted overlapping sub-regions is counted, and the ratio of this number to the total number of overlapping sub-regions is calculated to obtain the highlighted overlap ratio, which is then labeled as CS.

[0093] The mean deviation of the point cloud density of the overlapping sub-regions is obtained by subtracting the mean deviation of the point cloud density of the overlapping sub-regions. The mean deviation of the point cloud density of all overlapping sub-regions is summed and averaged to obtain the relative deviation of the point cloud density. The ratio of the relative deviation of the point cloud density to the mean deviation of the point cloud density is processed to obtain the value of the degree of overlap, and it is marked as CZ.

[0094] The ratio of prominent overlap (CS) and the value of prominent overlap (CZ) are processed using the following formula: The non-uniform point cloud density value DJ is obtained, where s1 and s2 are both preset scaling coefficients;

[0095] The point cloud distribution performance value is obtained by summing the point cloud density non-uniformity value DJ with the point cloud deviation ratio, and then labeled as DYB.

[0096] The obtained point cloud distribution representation value DYB and overlap degree CD are processed using the formula: The model reconstruction value CG is obtained, where a1 and a2 are preset scaling coefficients;

[0097] It should be noted that the model reconstruction value CG means that the model reconstruction value CG is calculated by the point cloud distribution representation value DYB and the overlap degree CD. The overlap degree reflects the degree of overlap between the spliced ​​segment track models, and the point cloud distribution representation value DYB reflects the point cloud distribution density of the overlapping area between the spliced ​​segment track models. The larger the model reconstruction value CG, the more accurate the spliced ​​segment track model is.

[0098] In some embodiments, the model reconstruction value CG is compared with the model reconstruction threshold;

[0099] If the model reconstruction value CG is greater than or equal to the model reconstruction threshold, it indicates that the overlap of the overlapping area is high and the point cloud distribution is uniform and the density is qualified, so no operation is performed.

[0100] If the model reconstruction value CG is less than the model reconstruction threshold, it indicates that the overlap of the overlapping area is low and the point cloud distribution is uneven and the density is unqualified. In this case, the two segmented track sections corresponding to the overlapping area are selected for model reconstruction.

[0101] Based on the selected segmented track, surface feature data of the selected segmented track is obtained using a cross-section scanner. Combined with the attitude and position information from the inertial navigation system, and using BIM technology, a 3D model of the selected segmented track is reconstructed. Figure 3 As shown (the left figure reflects the measured track entity, and the right figure reflects the track 3D model after modeling), the high-precision 3D model of the reconstructed segmented track is finally spliced ​​with the unreconstructed segmented track model and exported into a general format to obtain the track geometry model.

[0102] For example, surface feature data of the selected segmented track is obtained using a cross-section scanner, and combined with the attitude and position information of the inertial navigation system, specifically:

[0103] A cross-section scanner is used to scan the surface features of a track, acquiring surface feature data such as the shape, size, and contour of the track cross-section.

[0104] Inertial navigation systems can collect and process data such as attitude, velocity and acceleration of an object in motion in real time, providing accurate attitude and position information for estimating the object's position in space.

[0105] The data collected by the cross-section scanner and inertial navigation system are stitched together, registered and reconstructed to generate a three-dimensional model of the track;

[0106] It should be noted that by fusing data from the cross-section scanner and the inertial navigation system, data splicing, registration, and reconstruction can be achieved to generate a consistent and high-precision 3D model of the track. This data fusion not only improves upon the limitations of a single system, but also achieves more comprehensive and accurate monitoring by making full use of multi-source data. Efficient data fusion and registration are key steps to ensure the accuracy of the model and can better reflect the true condition of the track.

[0107] The technical solution of this embodiment is as follows: A three-dimensional model of the segmented track is established based on aerial photography information from UAVs and BIM technology, and preliminary splicing is performed. The overlapping areas and performance data of adjacent segmented track 3D models are obtained after splicing. Based on the processing and analysis of the performance data, segmented track models are selected. Surface feature data of the selected segmented track is obtained using a cross-section scanner. Combined with attitude and position information from an inertial navigation system, the three-dimensional model of the selected segmented track is reconstructed using BIM technology. The high-precision 3D model of the reconstructed segmented track is then finally spliced ​​with the unreconstructed segmented track model and exported in a general format to obtain the track geometry model. UAV aerial photography modeling provides a good foundation for macroscopic modeling, allowing for overall modeling of the track and surrounding facilities. Due to the high mobility of UAVs, this modeling process has high efficiency. Furthermore, the combination of a cross-section scanner and an inertial navigation system enables compensation and reconstruction of the track model, improving modeling accuracy. This has practical engineering significance for the modeling process of maglev track and is beneficial for achieving intelligent monitoring of the track.

[0108] Example 2

[0109] like Figure 1 As shown, based on Embodiment 1, the method for monitoring the digital twin status of maglev line tracks by integrating multi-dimensional information from air, rail, and ground, as described in this embodiment of the invention, includes:

[0110] Step 2: Based on the air-track-ground multi-dimensional information acquisition system, obtain holographic status data of the maglev line track. Based on the processing and analysis of the holographic status data, obtain the missing data values. Based on the missing data values, evaluate the holographic status data. Based on the qualified holographic status data, construct a database. Combined with the completed geometric model, use BIM technology to construct the behavior and rule model of the digital base of the maglev line track.

[0111] It should be noted that the air-rail-ground multi-dimensional information acquisition system consists of aerial data acquisition by UAVs, dynamic track data acquisition, and a fixed structural health monitoring system established by typical track beams;

[0112] The process of acquiring holographic state data includes:

[0113] A1. Technical means of high-altitude measurement: UAVs conduct regular automatic flight inspections to achieve rapid detection of the status information of line ancillary facilities such as cable trenches, switch houses, and substations;

[0114] A2, Technical means of dynamic track inspection: on-orbit machine vision inspection and structured light scanning. The on-orbit machine vision inspection system refers to the use of industrial cameras and edge computers as the main components, based on high-speed shooting and machine learning algorithms to identify the following information: structural seam recognition, foreign object intrusion in functional areas, stator surface wear, and substructure defects, and record the mileage information and anomaly type of the identified location; the structured light profile scanner is used to accurately scan the stator surface, and the stator surface deviation and fold angle are measured by point cloud processing to describe the geometric shape and position deviation.

[0115] The first example is, such as Figure 4 As shown, structured light scanning monitoring: Structured light contour scanning acquires information such as the shape, size, and topology of an object by performing high-precision measurement and reconstruction of the object's surface. This technology is based on the principle of structured light, using a light source to project onto the object's surface, and then using a camera to capture and analyze the projected image to ultimately form a three-dimensional model of the object's surface. It has advantages such as high scanning accuracy, fast scanning speed, and simple operation. The high-precision three-dimensional data provided by structured light scanning can be fused with data from other monitoring systems, such as inertial navigation systems, at a "microscopic" level. Through comprehensive data analysis, it enables comprehensive monitoring of the track condition.

[0116] The second example is, such as Figure 5 As shown, industrial camera inspection: by acquiring object images at high speed and using digital signal processing (DSP) chips for real-time image preprocessing and algorithm calculation, it can quickly and accurately analyze, classify and identify defects in objects, and generate statistical data and alarm information to feed back to operators or downstream processes; this inspection takes high-speed maglev tracks as the main inspection object, and the inspection items include: structural seam identification, sliding surface anomaly detection, etc.

[0117] A3. Select typical track beams to establish a fixed structural health monitoring system and build an all-weather fixed-point monitoring network for track structure and environmental conditions. The monitoring objects include: track beam end inclination angle, mid-span vibration response characteristics, surface temperature distribution, and environmental conditions.

[0118] For example, such as Figure 6 As shown, ground-based fixed-point monitoring in a fixed structural health monitoring system includes:

[0119] (1) Structural performance service status assessment: Through long-term monitoring and analysis of data such as acceleration, strain, and inclination of the track structure, the performance status of the track during service is assessed. The main contents of the assessment include identifying potential structural damage, fatigue, and deformation to ensure the safety and reliability of the track;

[0120] (2) Calculation of train passing time, direction of travel and timetable of line: Analyze the time when the train passes through a specific monitoring point using sensor data, and calculate the direction of travel and speed of the train by combining dynamic response data on the track; further, by summarizing this information, the timetable of line can be constructed and optimized to improve operational efficiency.

[0121] (3) Vibration path study of vehicle-induced vibration: By studying the track vibration path caused by train operation through spatiotemporal analysis of vibration data, the vibration source and its propagation path can be identified. The research helps to understand the impact of train operation on the track and its surrounding environment and to take effective vibration reduction measures.

[0122] (4) Data fusion of multiple physical quantities: Multi-source heterogeneous data from different sensors are fused to obtain more comprehensive and accurate track monitoring information. Information fusion algorithms can effectively integrate sampling data from different sensors, improve the accuracy and consistency of the overall data, and enhance the real-time monitoring capability of track status;

[0123] (5) Interaction between train and environment: Analyze the impact of train operation on the environment (including parameters such as noise and electromagnetic field strength), and study the impact of environmental factors such as temperature, humidity and wind speed on track structure and train operation. Based on these studies, train operation and track maintenance strategies can be further optimized to improve the sustainability of the system.

[0124] In some embodiments, a collection period is set, the collection period is divided into several time analysis nodes, holographic state data at the time analysis nodes is obtained, and the holographic state data at the time analysis nodes is compared with standard holographic state data.

[0125] If the total amount of data contained in the holographic state data at the time analysis node is not equal to the total amount of data contained in the standard holographic state data, then the time analysis node will be marked as an abnormal acquisition node.

[0126] If the total amount of data contained in the holographic state data at the time analysis node is equal to the total amount of data contained in the standard holographic state data, then no processing is performed;

[0127] It should be noted that the standard holographic state data refers to the holographic state data of the maglev line track that needs to be collected, which is set according to the actual needs of the behavior and rule model.

[0128] The number of abnormal data collection nodes is counted, and the ratio of this number to the number of time analysis nodes is calculated to obtain the abnormal data collection ratio, which is then labeled as SL.

[0129] The difference between the total amount of data contained in the holographic state data at the abnormal acquisition node and the total amount of data contained in the standard holographic state data is processed to obtain the data missing amount at the abnormal acquisition node. The data missing amounts at all abnormal acquisition nodes are summed and averaged to obtain the average data missing amount. The ratio of the average data missing amount to the total amount of data contained in the standard holographic state data is processed to obtain the abnormal acquisition missing ratio, which is marked as QS.

[0130] The obtained outlier count ratio (SL) and outlier missing ratio (QS) are processed using the following formula: The missing data value QP is obtained, where z1 and z2 are both preset scaling coefficients;

[0131] In some embodiments, the data missing value QP is compared with the data missing threshold;

[0132] If the missing data value QP is greater than or equal to the missing data threshold, it indicates that the holographic state data collected during the acquisition period is highly missing and is unqualified.

[0133] If the missing data value QP is less than the missing data threshold, it indicates that the degree of missing data in the holographic state collected during the collection period is low and acceptable.

[0134] It should be noted that, based on the inadequacy of the data, interpolation can be used to fill in the missing data;

[0135] Based on qualified holographic state data, it is imported into the track geometry model constructed using BIM technology to complete the establishment of the maglev line system behavior model;

[0136] Meanwhile, the collected holographic state data was processed and analyzed using Matlab data analysis tools, including: identifying patterns, trends and outliers in the holographic state data; extracting feature variables useful for building a rule model using Matlab data analysis tools; and constructing a rule model for the maglev line based on the data analysis results.

[0137] Step 3: Based on qualified holographic state data, use the Kalman filter fusion algorithm to fuse the information of homogeneous and heterogeneous holographic state data within the holographic state data, so as to realize diversified intelligent monitoring of the maglev line track.

[0138] The technical solution of this invention is as follows: Based on a multi-dimensional information acquisition system of air, rail, and ground, holographic state data of maglev line tracks is obtained. Based on the processing and analysis of the holographic state data, missing data values ​​are obtained. The holographic state data is evaluated based on these missing values. A database is constructed based on the qualified holographic state data. Combined with the completed geometric model, a behavior and rule model of the digital base of the maglev line track is constructed using BIM technology. Based on the qualified holographic state data, a Kalman filter fusion algorithm is used to fuse isomorphic and heterogeneous holographic state data. This invention realizes the construction of a behavior and rule model of the digital base of the maglev line track, completes the monitoring of the maglev line track, and fuses its diverse holographic state data, achieving diversified intelligent monitoring of the maglev line track.

[0139] Example 3

[0140] like Figure 1 As shown, based on Embodiments 1 and 2, the method for monitoring the digital twin status of maglev line tracks by integrating multi-dimensional information from air, rail, and ground, as described in this embodiment of the invention, includes:

[0141] Step 4: Based on the collected qualified holographic status data and the geometric model, behavior and rule model of the track, construct a smart monitoring platform for maglev track.

[0142] Specifically, it includes:

[0143] H1, using MATLAB or Python data analysis software to analyze the holographic state data of the maglev line track system, specifically as follows:

[0144] Track geometry: including track dimensions, alignment, levelness, and elevation difference;

[0145] Structural health status: Monitoring parameters such as stress, strain, vibration, and temperature of structural components such as track beams, supports, and connectors;

[0146] Environmental parameters: such as temperature, humidity, wind speed, and rainfall;

[0147] Equipment status: The operating status and performance parameters of equipment such as signal systems, power supply systems, and communication systems;

[0148] H2 utilizes decentralized edge computing technology to perform preliminary data preprocessing at the device end (i.e., the sensor or data acquisition device) to reduce data transmission volume, improve the efficiency of cloud-to-device information transmission, and reduce cloud processing pressure. Specifically:

[0149] Gross error assessment: At the data acquisition point, preliminary screening is performed to remove obviously erroneous data;

[0150] Data cleaning: removing duplicate data, filling in missing values, handling outliers, etc., to improve data quality;

[0151] H3 uses cloud computing platforms (such as AWS, Azure, Alibaba Cloud, etc.) as its foundation, imports the railway track model, and completes the construction of the intelligent monitoring platform, specifically including:

[0152] Model demonstration: Import the geometric model, behavior and rule model of the track into the cloud platform to demonstrate the structure and status of the maglev track system;

[0153] Data Download: Provides a data download interface, allowing users to download raw or processed data as needed;

[0154] Database Management: Design and implement an efficient database management system for storing and managing various types of data for the maglev track system;

[0155] Login and Access Control: Provides user login and access control functions to ensure system security and data confidentiality.

[0156] The technical solution of this invention is as follows: Based on the collected qualified holographic status data and the geometric model, behavior and rule model of the track, a smart monitoring platform for maglev line track is constructed. This invention realizes intelligent monitoring of the track, improves the convenience and efficiency of maglev line track monitoring, and firstly, the invention collects and processes information from multiple dimensions (air, track and ground) to construct a track model and realize detailed and comprehensive monitoring of the track status. Then, through data processing, the basic data is connected in an orderly manner to ensure the accuracy and scientific nature of the data, such as data anomaly analysis. Finally, through a digital twin platform, the fusion and analysis results of multi-dimensional data are organically combined with the final monitoring decision to form a complete smart monitoring system.

[0157] Example 4

[0158] like Figure 7 As shown in the embodiment of the present invention, the maglev line track digital twin status monitoring system integrating multi-dimensional information on air, rail, and ground includes:

[0159] The route model construction module: Based on aerial photography information from UAVs and BIM technology, a 3D model of the segmented route track is established and initially stitched together. The overlapping areas and their representational data between adjacent segmented route track 3D models are obtained, including the area of ​​the overlapping area and the number of point cloud points. Based on the processing and analysis of the representational data, the segmented route tracks are selected. Surface feature data of the selected segmented route tracks are obtained using a cross-section scanner, combined with attitude and position information from the inertial navigation system, and the 3D model of the selected segmented route tracks is reconstructed using BIM technology. Finally, the high-precision 3D model of the reconstructed segmented route track is stitched together with the unreconstructed segmented route track model and exported to a universal format to obtain the route track geometric model.

[0160] Behavior and rule model construction module: Based on the air-track-ground multi-dimensional information acquisition system, obtain holographic state data of the maglev line track. Based on the processing and analysis of the holographic state data, obtain the missing data values. Based on the missing data values, evaluate the holographic state data. Based on the qualified holographic state data, construct a database. Combined with the completed geometric model, use BIM technology to construct the behavior and rule model of the digital base of the maglev line track.

[0161] Multi-source data fusion module: Based on qualified holographic state data, the Kalman filter fusion algorithm is used to fuse information between homogeneous and heterogeneous holographic state data within the holographic state data.

[0162] Intelligent monitoring platform construction module: Based on qualified holographic status data and the geometric, behavioral and rule models of the track, an intelligent monitoring platform for maglev lines is constructed.

[0163] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely illustrative of the principles of the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the present invention as claimed. The scope of protection of the present invention is defined by the appended claims and their equivalents.

Claims

1. A method for monitoring the digital twin status of maglev line tracks by integrating multi-dimensional information from air, rail, and ground, characterized by: include: Based on aerial photography information from UAVs and BIM technology, a 3D model of the segmented track is established and initially stitched together. The overlapping area between adjacent segmented track 3D models and the performance data of the overlapping area are obtained. The performance data includes the area of ​​the overlapping area and the number of point cloud points. Based on the processing and analysis of the performance data, the segmented track is selected. The process of selecting the segmented track is as follows: The obtained point cloud distribution representation value DYB and overlap degree CD are processed using the formula: The model reconstruction value CG is obtained, where a1 and a2 are preset scaling coefficients; Compare the model reconstruction value CG with the model reconstruction threshold; If the model reconstruction value CG is greater than or equal to the model reconstruction threshold, it indicates that the overlap of the overlapping area is high and the point cloud distribution is uniform and the density is qualified, so no operation is performed. If the model reconstruction value CG is less than the model reconstruction threshold, it indicates that the overlap of the overlapping area is low and the point cloud distribution is uneven and the density is unqualified. In this case, the two segmented track sections corresponding to the overlapping area are selected for model reconstruction. The surface feature data of the selected segmented track is obtained using a cross-section scanner. Combined with the attitude and position information of the inertial navigation system, the three-dimensional model of the selected segmented track is reconstructed using BIM technology. The high-precision three-dimensional model of the reconstructed segmented track is then stitched together with the unreconstructed segmented track model and exported in a general format to obtain the geometric model of the track. Based on the multi-dimensional information acquisition system of air-rail-ground, holographic state data of maglev line track is obtained. Based on the processing and analysis of holographic state data, missing data values ​​are obtained. Based on the missing data values, the holographic state data is evaluated. Based on the qualified holographic state data, a database is constructed. Combined with the completed geometric model, the behavior and rule model of the digital base of maglev line track is constructed using BIM technology. Based on qualified holographic state data, the Kalman filter fusion algorithm is used to fuse information between isomorphic and heterogeneous holographic state data within the holographic state data. A smart monitoring platform for maglev lines is constructed by integrating qualified holographic status data with geometric, behavioral, and rule models of the track.

2. The method for monitoring the digital twin status of maglev line tracks by integrating multi-dimensional information from air, rail, and ground, as described in claim 1, is characterized in that: The overlap CD is obtained as follows: Key feature points are extracted from the model using feature point detection and description algorithms, and similar feature points in adjacent models are found using feature point matching algorithms. Overlapping regions between models are obtained through feature point indicators. Obtain the area of ​​the overlapping region and the area of ​​the minimum bounding rectangle corresponding to the model. Calculate the ratio between the area of ​​the overlapping region and the area of ​​the minimum bounding rectangle corresponding to the model to obtain the overlap degree, and mark it as CD. The minimum bounding rectangle refers to the smallest rectangle that can enclose all target points, whose width and height are determined by the maximum and minimum values ​​of the target points.

3. The method for monitoring the digital twin status of maglev line tracks by integrating multi-dimensional information from air, rail, and ground, as described in claim 1, is characterized in that: The point cloud distribution representation value DYB is obtained as follows: The ratio of prominent overlap (CS) and the value of prominent overlap (CZ) are processed using the following formula: The non-uniform point cloud density value DJ is obtained, where s1 and s2 are both preset scaling coefficients; The point cloud distribution performance value is obtained by summing the point cloud density non-uniformity value DJ with the point cloud deviation ratio, and then labeled as DYB. The point cloud deviation ratio is obtained as follows: Point cloud analysis is performed on the overlapping area using point cloud processing software to obtain the number of point clouds in the overlapping area. The point cloud density of the overlapping area is obtained by comparing the number of point clouds in the overlapping area with the area of ​​the overlapping area. The difference between the point cloud density of the overlapping area and the set point cloud density of the overlapping area is calculated, and the absolute value of the difference is taken to obtain the point cloud density difference of the overlapping area. The point cloud density difference of the overlapping area is then compared with the set point cloud density of the overlapping area to obtain the point cloud deviation ratio.

4. The method for monitoring the digital twin status of maglev line tracks by integrating multi-dimensional information from air, rail, and ground, as described in claim 3, is characterized in that: The method for obtaining the prominent overlap ratio CS is as follows: The overlapping region is divided into several overlapping sub-regions with equal areas. The number of point clouds in each overlapping sub-region is obtained and then compared with the area of ​​the overlapping sub-region to obtain the point cloud density of the overlapping sub-region. The point cloud density of all overlapping sub-regions is summed and averaged to obtain the average point cloud density. The point cloud density of each overlapping sub-region is then compared with the average point cloud density, and the absolute value of the difference is taken to obtain the point cloud density deviation of the overlapping sub-region. The point cloud density deviation of all overlapping sub-regions is summed and averaged to obtain the average point cloud density deviation. The point cloud density deviation of the overlapping sub-region is then compared with the average point cloud density deviation. If the point cloud density deviation of an overlapping sub-region is greater than the mean point cloud density deviation, then the overlapping sub-region is marked as a prominent overlapping sub-region. If the point cloud density deviation of the overlapping sub-region is less than or equal to the mean point cloud density deviation, then the overlapping sub-region is marked as a non-prominent overlapping sub-region. The number of highlighted overlapping sub-regions is counted, and the ratio of this number to the total number of overlapping sub-regions is calculated to obtain the highlighted overlap ratio, which is then labeled as CS.

5. The method for monitoring the digital twin status of maglev line tracks by integrating multi-dimensional information from air, rail, and ground, as described in claim 3, is characterized in that: The method for obtaining the degree of overlap CZ is as follows: The mean deviation of the point cloud density between the overlapping sub-regions and the mean point cloud density is processed to obtain the mean deviation of the point cloud density of the overlapping sub-regions. The mean deviation of the point cloud density of all overlapping sub-regions is summed and averaged to obtain the relative deviation of the point cloud density. The ratio of the relative deviation of the point cloud density to the mean point cloud density deviation is processed to obtain the degree of overlap, and it is labeled as CZ.

6. The method for monitoring the digital twin status of maglev line tracks by integrating multi-dimensional information from air, rail, and ground, as described in claim 1, is characterized in that: The process of acquiring the holographic state data includes: A1. Technical means of high-altitude measurement: UAVs conduct regular automatic flight inspections to achieve rapid detection of the status information of cable trenches, switch houses, and substation line auxiliary facilities; A2, Technical means of dynamic track inspection: on-orbit machine vision inspection and structured light scanning. The on-orbit machine vision inspection system refers to the use of industrial cameras and edge computers as the main components, based on high-speed shooting and machine learning algorithms to identify the following information: structural seam recognition, foreign object intrusion in functional areas, stator surface wear, and substructure defects, and record the mileage information and anomaly type of the identified location; the structured light profile scanner is used to accurately scan the stator surface, and the stator surface deviation and fold angle are measured by point cloud processing to describe the geometric shape and position deviation. A3. Select typical track beams to establish a fixed structural health monitoring system and build an all-weather fixed-point monitoring network for track structure and environmental conditions. The monitoring objects include: track beam end inclination angle, mid-span vibration response characteristics, surface temperature distribution, and environmental conditions.

7. The method for monitoring the digital twin status of maglev line tracks by integrating multi-dimensional information from air, rail, and ground, as described in claim 1, is characterized in that: The missing data value QP is obtained as follows: Set the acquisition period, divide the acquisition period into several time analysis nodes, acquire the holographic state data at the time analysis nodes, and compare the holographic state data at the time analysis nodes with the standard holographic state data. If the total amount of data contained in the holographic state data at the time analysis node is not equal to the total amount of data contained in the standard holographic state data, then the time analysis node will be marked as an abnormal acquisition node. If the total amount of data contained in the holographic state data at the time analysis node is equal to the total amount of data contained in the standard holographic state data, then no processing is performed; The number of abnormal data collection nodes is counted, and the ratio of this number to the number of time analysis nodes is calculated to obtain the abnormal data collection ratio, which is then labeled as SL. The difference between the total amount of data contained in the holographic state data at the abnormal acquisition node and the total amount of data contained in the standard holographic state data is processed to obtain the data missing amount at the abnormal acquisition node. The data missing amounts at all abnormal acquisition nodes are summed and averaged to obtain the average data missing amount. The ratio of the average data missing amount to the total amount of data contained in the standard holographic state data is processed to obtain the abnormal acquisition missing ratio, which is marked as QS. The obtained outlier count ratio (SL) and outlier missing ratio (QS) are processed using the following formula: The missing data value QP is obtained, where z1 and z2 are preset scaling coefficients.

8. The method for monitoring the digital twin status of maglev line tracks by integrating multi-dimensional information from air, rail, and ground, as described in claim 1, is characterized in that: The process of evaluating the holographic state data based on the missing data value is as follows: comparing the missing data value QP with the missing data threshold. If the missing data value QP is greater than or equal to the missing data threshold, it indicates that the holographic state data collected during the acquisition period is highly missing and is unqualified. If the missing data value QP is less than the missing data threshold, it indicates that the degree of missing data in the holographic state collected during the acquisition period is low and acceptable.

9. The method for monitoring the digital twin status of maglev line tracks by integrating multi-dimensional information from air, rail, and ground, as described in claim 1, is characterized in that: The specific process of integrating and constructing the intelligent monitoring platform for maglev line tracks is as follows: H1, using MATLAB or Python data analysis software to analyze the holographic state data of the maglev line track system, specifically as follows: Track geometry: including track dimensions, alignment, levelness, and elevation difference; Structural health status: Monitor the stress, strain, vibration, and temperature parameters of the track beams, supports, and connecting structural components; Environmental parameters: such as temperature, humidity, wind speed, and rainfall; Equipment status: Operating status and performance parameters of signal system, power supply system, and communication system equipment; H2 utilizes decentralized edge computing technology to perform preliminary data preprocessing at the device level, thereby reducing data transmission volume, improving the efficiency of information transmission between the cloud and the device, and reducing the processing pressure on the cloud. Specifically: Gross error assessment: At the data acquisition point, preliminary screening is performed to remove obviously erroneous data; Data cleaning: removing duplicate data, filling in missing values, handling outliers, and improving data quality; H3 uses a cloud computing platform as its foundation, imports the railway track model, and completes the construction of a smart monitoring platform, specifically including: Model demonstration: Import the geometric model, behavior and rule model of the track into the cloud platform to demonstrate the structure and status of the maglev track system; Data Download: Provides a data download interface, allowing users to download raw or processed data as needed; Database Management: Design and implement an efficient database management system for storing and managing various types of data for the maglev track system; Login and Access Control: Provides user login and access control functions to ensure system security and data confidentiality.

Images