Digital twinborn simulation system for sleeper maintenance period

Abstract

The invention relates to the technical field of digital twinborn simulation, in particular to a digital twinborn simulation system for a sleeper maintenance period, which comprises a multi-source data integration module for integrating point cloud data of a laser scanner and a photogrammetric image sequence of an unmanned aerial vehicle, extracting steel rail profile point position coordinates and sleeper position geometric parameters of a sleeper, and generating a digital twinborn simulation result; and generating an integrated structure attribute set. According to the invention, by integrating the point cloud data of the laser scanner and the photogrammetric image sequence of the unmanned aerial vehicle, the space coordinates of the steel rail profile and the geometric parameters of the sleeper are obtained, and a unified space coordinate system is constructed, so that the accuracy and stability of data matching are improved; and the displacement vector and attitude difference between the actually measured data and the digital model are taken as an adjustment basis, so that the adjustment operation of the steel rail node coordinate and the sleeper rotation matrix is more in line with the actual track facility state, the consistency of the simulation geometric structure and the actual track structure is realized, and the credibility of model simulation analysis is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of digital twin simulation, and particularly to a digital twin simulation system for the maintenance cycle of sleeper rails. Background Art

[0002] In the technical field of digital twin simulation, physical entities or systems in the real world are digitally mirrored in the virtual space through information collection, data fusion, and dynamic modeling to create a digital mirror model with a complete mapping relationship, enabling real-time monitoring, simulation analysis, and prediction of physical systems. However, existing technologies generally adopt a single model fineness, showing deficiencies in regional computing resource allocation, which easily leads to a reduction in simulation efficiency and an increase in additional computing overhead. Therefore, improvements are needed. Summary of the Invention

[0003] The purpose of the present invention is to address the shortcomings in the prior art and propose a digital twin simulation system for the maintenance cycle of sleeper rails.

[0004] To achieve the above purpose, the present invention adopts the following technical solutions: The digital twin simulation system for the maintenance cycle of sleeper rails includes: A multi-source data integration module that integrates the point cloud data of a laser scanner and the image sequence of UAV photogrammetry, extracts the point coordinates of the rail profile of the sleeper rail and the geometric parameters of the sleeper position, and generates an integrated structure attribute set; A geometric consistency calibration module that, based on the preset geometric form of the digital twin model and the integrated structure attribute set, calculates the spatial displacement vector between the reference points of the corresponding rail components in the digital twin model and the measured coordinates, obtains a geometric deviation metric element, and adjusts the coordinates of the corresponding rail nodes and the rotation matrix of the sleeper components in the digital twin model according to the geometric deviation metric element to map the geometric form change and generate a calibrated geometric representation; A topological structure evolution module that retrieves the calibrated geometric representation and, in combination with the maintenance records during the maintenance cycle of the sleeper rail, determines the changes in the component connection relationship caused by sleeper replacement or fastener addition or subtraction, forms a connection change instruction set, and reconstructs the constraints between the rail and the sleeper according to the connection change instruction set to construct an evolved topological network diagram; An adaptive collaborative simulation module that receives the evolved topological network diagram and the simulation analysis requirements of the maintenance cycle of the sleeper rail, selects a matching track unit model according to the identified stress area, combines it into a hybrid fidelity model configuration, and performs simulation operations based on the hybrid fidelity model configuration to obtain an adaptive simulation output result.

[0005] Preferably, the steps for obtaining the integrated structure attribute set are as follows: Based on the track point cloud data collected by the laser scanner and the image sequences obtained by the UAV aerial photography, extract the three-dimensional coordinate of the rail edge contour points and the geometric position parameters of the sleeper cross-section center point in each frame. After uniformly projecting them into the WGS-84 coordinate system, align all the acquisition timestamps, and form a set of combined feature point pairs by indexing and matching the corresponding frame sequences one by one to obtain an aligned list of combined feature point pairs; Calculate the rail profile curve registration factor according to the aligned list of combined feature point pairs; According to the rail profile curve registration factor and the three-dimensional offset value of each pair of feature points, screen the point pairs with offset values greater than the rail profile curve registration factor, recalculate the remapping vector in the global coordinate system and perform coordinate substitution. At the same time, perform coordinate fitting and correction on the geometric parameters of the sleeper cross-section center point, and combine the remapped point coordinates and the corrected geometric parameters to generate an integrated structure attribute set.

[0006] Preferably, the steps for obtaining the geometric deviation metric element are as follows: Based on the standard geometric shapes of the rail and the sleeper preset in the digital twin model, call the rail edge contour point coordinates and the geometric parameters of the sleeper cross-section center point in the integrated structure attribute set, match the corresponding rail component reference points and the sleeper component center points in the digital twin model to form a one-to-one corresponding spatial point pair between the model and the measured points; Based on the coordinates of each pair of rail component reference points and the measured rail edge contour points in the spatial point pair, calculate the coordinate differences of this point pair along the X, Y, and Z axes in the three-dimensional space one by one, synthesize a complete three-dimensional space displacement vector and record the vector length. At the same time, calculate the average spatial displacement vector length of the point pair set to generate a rail spatial displacement vector set; Based on the geometric parameters of each pair of sleeper component center points and the measured sleeper cross-section center points in the spatial point pair, extract the Euler angle difference values of the rotational attitude changes between the two points, and record the attitude angle differences on the pitch angle, yaw angle, and roll angle respectively to form a sleeper attitude angle difference set. Combine the rail spatial displacement vector set and the sleeper attitude angle difference set to obtain the geometric deviation metric element.

[0007] Preferably, the steps for obtaining the calibrated geometric representation are as follows: According to the geometric deviation metric element, extract each spatial displacement vector in the rail spatial displacement vector set, and respectively analyze the displacement components of this spatial displacement vector on the X, Y, and Z coordinate axes. For each corresponding rail node coordinate in the digital twin model, add the numerical values of the analyzed displacement components on each axis to the original coordinate value of the node to form an adjusted rail node coordinate sequence; Based on the adjusted rail node coordinate sequence, call the set of sleeper attitude angle differences in the geometric deviation metric element, parse the pitch angle difference, yaw angle difference, and roll angle difference of each sleeper in the set one by one, recalculate the rotation matrix parameters of each sleeper component, and generate a corrected sleeper rotation matrix sequence; Based on the adjusted rail node coordinate sequence and the corrected sleeper rotation matrix sequence, apply the rail node coordinate sequence and the sleeper rotation matrix sequence to the original geometric configuration of the rails and sleepers in the digital twin model one by one, and perform a three-dimensional shape mapping update again to form a calibrated geometric representation.

[0008] Preferably, the steps for obtaining the connection change instruction set are as follows: Retrieve the rail node coordinates and sleeper component rotation matrices in the calibrated geometric representation, and item by item identify the adjacency information and connection type identifiers in the original connection relationships between each node and each sleeper. Combine the text content of the maintenance records during the rail-sleeper maintenance period, extract all record statements with keywords such as "replacement", "removal", and "new addition", and parse the associated sleeper numbers and fastener numbers to form an initial list of structural connection changes; Based on the initial list of structural connection changes, compare the parsed sleeper numbers and fastener numbers with the current connection status in the calibrated geometric representation according to the time period and spatial position interval corresponding to each sleeper number. If there is a sleeper number in the list but not in the calibrated geometric representation, it is determined as a newly added sleeper connection. If there is a fastener number in the calibrated geometric representation but not in the list, it is determined as a fastener removal, and a list of component connection relationship changes is generated; According to the list of component connection relationship changes, construct a standardized instruction text format for all connection relationships marked as newly added or removed, and sequentially supplement the instruction type, target number, operation node index, and influence range identifier to generate a connection change instruction set.

[0009] Preferably, the steps for obtaining the evolutionary topology network diagram are as follows: Based on the connection change instruction set, parse the target number, operation node index, and influence range identifier in each connection change instruction in sequence, call the rail node coordinates and sleeper component indices in the calibrated geometric representation, and determine whether the target number corresponds to the existing node set. If it is a new addition instruction, add this number to the node set and construct an initial connection entry. If it is a removal instruction, delete all connection entries associated with this number to obtain a structural connection reconstruction node table; According to the structural connection reconstruction node table, scan the connection status between each rail node coordinate and the corresponding sleeper component, extract all node pairs with physical connection relationships, and calculate the connection orientation attributes according to the rotation matrix status of the sleeper components to form a constraint relationship reconstruction entry table; Reconstruct the entry table based on the above-mentioned constraint relationship, establish an adjacency list structure with the node number as the main index one by one, embed the connection status, connection direction, and connection label as edge attributes into the corresponding adjacency entries, and simultaneously construct an edge index table to generate an evolutionary topology network diagram.

[0010] Preferably, the steps for obtaining the hybrid fidelity model configuration are as follows: Based on the node structure relationship in the evolutionary topology network diagram, extract the set of connecting edges between all node pairs, and identify the node numbers and spatial distribution positions associated with each connecting edge one by one. Match the structural fatigue monitoring parameter table given in the simulation analysis requirements for sleeper track maintenance cycles, screen the set of regional edges where the connecting edges fall into the fatigue risk identification interval, and generate a stress region boundary index list; Based on the stress region boundary index list, calculate the track element model adaptation factor for each stress region; Based on the track element model adaptation factor corresponding to each stress region, compare the track element model adaptation factor with the adaptation value benchmark threshold item by item. If the track element model adaptation factor corresponding to the region is greater than or equal to the adaptation value benchmark threshold, designate the region as a track element model with a complete parameter configuration. Otherwise, designate it as a track element model with a parameter compression configuration. Integrate the two types of track element models into the hybrid fidelity model configuration of the overall track system.

[0011] Preferably, the steps for obtaining the adaptive simulation output result are as follows: Based on each region delimited by the hybrid fidelity model configuration, call the track element models with complete parameter configurations or parameter compression configurations region by region, respectively load the current load conditions and material parameters of each track element model in the region, perform dynamic stress analysis and calculation, and obtain the real-time data sequence of node stresses in each region in real time to generate a real-time data set of regional node stresses; According to the real-time data set of regional node stresses, calculate the difference in stress values within consecutive time steps in each region and divide it by the corresponding time interval to obtain the stress gradient change rate in each region. Compare the calculated stress gradient change rate with the preset stress gradient change rate threshold region by region, identify the abnormal regions where the change rate exceeds the preset limit, and form an abnormal stress region identification list; Based on the abnormal stress region identification list, for each abnormal region in the list, upgrade the track element model originally configured with a parameter compression configuration to a complete parameter configuration, refine the parameters of the track element model originally configured with a complete parameter configuration, and update the stress data again to generate the adaptive simulation output result.

[0012] Compared with the prior art, the advantages and positive effects of the present invention are as follows: In the present invention, by integrating the point cloud data of a laser scanner and the image sequence of unmanned aerial vehicle photogrammetry, the spatial coordinates of the rail profile and the geometric parameters of the sleepers are obtained, and a unified spatial coordinate system is constructed, which improves the accuracy and stability of data matching; and taking the displacement vector and attitude difference between the measured data and the digital model as the adjustment basis, the adjustment operations of the rail node coordinates and the sleeper rotation matrix are made more in line with the actual state of the track facilities, realizing the consistency between the simulated geometric structure and the actual track structure, and improving the credibility of the model simulation analysis; then, according to the actual maintenance records of the maintenance cycle, the component connection changes caused by sleeper replacement and fastener addition and subtraction are identified, and the structural constraint relationship between the rail and the sleeper is reconstructed in time, so that the topological network can reflect the actual state changes of the facilities in real time, improving the dynamic evolution ability of the topological model; at the same time, according to the identified stress area, the model granularity and fidelity are adjusted, the complexity and computing load of the simulation model are controlled, the computing efficiency is improved while ensuring the simulation accuracy, and the simulation resource requirements are reduced. Brief Description of the Drawings

[0013] Figure 1 It is a system flow chart of the present invention. Detailed Embodiments

[0014] In order to make the objectives, technical solutions and advantages of the present invention clearer, the present invention will be further described in detail below with reference to the drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0015] Please refer to Figure 1 , the present invention provides a technical solution: a digital twin simulation system for the sleepers and rails maintenance cycle includes: A multi-source data integration module, which integrates the point cloud data of a laser scanner and the image sequence of unmanned aerial vehicle photogrammetry, extracts the point coordinates of the rail profile of the sleepers and rails and the geometric parameters of the sleeper positions, and generates an integrated structure attribute set; A geometric consistency calibration module, based on the preset geometric form of the digital twin model and the integrated structure attribute set, calculates the spatial displacement vector between the reference points of the corresponding rail components in the digital twin model and the measured coordinates, obtains a geometric deviation metric element, and adjusts the corresponding rail node coordinates and the rotation matrix of the sleeper components in the digital twin model according to the geometric deviation metric element, maps the geometric form changes, and generates a calibrated geometric representation; A topological structure evolution module, retrieves the calibrated geometric representation, and combines the maintenance records during the sleepers and rails maintenance cycle to determine the changes in the component connection relationships caused by sleeper replacement or fastener addition and subtraction, forms a connection change instruction set, and reconstructs the constraints between the rails and the sleepers according to the connection change instruction set to construct an evolving topological network diagram; The adaptive co-simulation module receives the evolved topological network diagram and the simulation analysis requirements for the maintenance cycle of the rail sleeper. According to the identified stress areas, it selects a matching track element model, combines them into a hybrid fidelity model configuration, and performs simulation operations based on the hybrid fidelity model configuration to obtain the adaptive simulation output results.

[0016] The steps for obtaining the integrated structure attribute set are as follows: Based on the rail point cloud data collected by the laser scanner and the image sequences obtained by the UAV aerial photography, the three-dimensional coordinates of the rail edge contour points and the geometric position parameters of the center points of the rail sleeper cross-sections in each frame are extracted. After uniformly projecting them into the WGS-84 coordinate system, all the acquisition timestamps are aligned. By indexing and matching the corresponding frame sequences one by one, a set of joint feature point pairs is formed, and a list of aligned joint feature point pairs is obtained. According to the list of aligned joint feature point pairs, calculate the rail profile curve registration factor, and the calculation formula is: ; where, is the rail profile curve registration factor, is the number of joint feature points, is the three-dimensional coordinate of the th point in the laser scanner, is the three-dimensional coordinate of the corresponding th point in the image measurement sequence; According to the rail profile curve registration factor and the three-dimensional offset value of each pair of feature points, select the point pairs with the offset value greater than the rail profile curve registration factor, recalculate the remapping vector in the global coordinate system and perform coordinate replacement. At the same time, perform coordinate fitting and correction on the geometric parameters of the center points of the rail sleeper cross-sections, and combine the remapped point coordinates and the corrected geometric parameters to generate the integrated structure attribute set.

[0017] Specifically, based on the track point cloud data collected by the laser scanner and the image sequences obtained from the UAV aerial photography, the laser point cloud data is first preprocessed, including removing the noise points and outliers generated by non-track objects (such as vegetation and debris) in the track environment, and excluding the points whose distances from neighboring points exceed three standard deviations. Then, the UAV image sequences are subjected to distortion correction and feature extraction. The internal and external parameters of the camera are used to geometrically correct each frame of the image, eliminating lens distortion and projection deformation, and the scale-invariant feature transform algorithm, for example, is used to extract the stable feature points in the image. Then, for the extraction of the three-dimensional coordinates of the rail edge contour points, in the laser point cloud data, the rail area is identified, and then the edge detection and fitting of the rail cross-section point cloud are carried out to obtain the edge points of the rail profile and record their three-dimensional coordinates. For the UAV images, the rail area is identified through a deep learning object detection model (such as YOLOv5), and then the rail edge is extracted using an edge detection operator (such as the Canny operator), and its three-dimensional coordinates are restored by combining the photogrammetry principle and the ground control points. For the extraction of the geometric position parameters of the center point of the sleeper cross-section, in the laser point cloud, the sleeper point cloud cluster is identified and its geometric center is calculated. In the UAV images, the sleeper is identified through template matching or object detection, and its image center point is calculated, and then its three-dimensional position is restored by triangulation or structured light projection. Subsequently, the coordinates extracted from the laser point cloud (usually based on the scanner's own coordinate system or the local engineering coordinate system) and the coordinates extracted from the UAV images (usually based on the image coordinate system or the geographic coordinate system of the GPS initial positioning) are uniformly projected onto the WGS-84 coordinate system. This process requires the GPS / IMU data recorded by each data acquisition device and the preset ground control points, and is converted through the coordinate transformation seven-parameter model (Bursa model). Then, the timestamps of all the collected data are aligned to ensure that the laser scan data frames and the UAV image frames are strictly corresponding in time. If the clocks of the acquisition devices are not precisely synchronized, the synchronization events that can be clearly identified in both data sources (such as the moment when a specific train passes) are identified or the clock signal is used for post-synchronization calibration. Finally, according to the aligned timestamps, the laser scan point cloud frames and the UAV image frames within the same moment or a very short time interval (such as 0.05 seconds) are associated through the index number. In these associated frames, the corresponding feature points from the two data sources are matched based on spatial proximity and feature similarity (such as the previously extracted rail edge points or sleeper corner points), forming a joint feature point pair containing the laser point coordinates and the corresponding image measurement point coordinates, and they are compiled into a list to obtain the aligned joint feature point pair list.

[0018] Formula: , the advantage of the formula is that by calculating the average three-dimensional Euclidean distance between the corresponding feature points in the laser scan data and the UAV image measurement data, it provides a quantitative evaluation index for the spatial registration quality of multi-source data, that is, the rail profile curve registration factor , which can judge the geometric consistency degree of the fusion of two data sources, provides a reliable basis for subsequent data screening and model calibration, and ensures the accuracy of the geometric shape of the digital twin model. Specifically, by the average distance of the joint feature points to measure the overall registration effect, avoiding the excessive influence of extreme deviations of single or a few point pairs on the overall evaluation; is the number of joint feature points, referring to the total number of pairs of homologous feature points successfully matched in the laser scanning data and the unmanned aerial vehicle image measurement data. This parameter directly comes from the statistical count of the previous step "obtaining the list of aligned joint feature point pairs". For example, if this list contains 1500 pairs of successfully matched feature points, then .

[0019] is the three-dimensional coordinates of the th point in the laser scanner, indicating that in the "list of aligned joint feature point pairs", the eastward coordinate, northward coordinate, and elevation value of the point from the laser scanner in the th pair of joint feature points in the WGS-84 coordinate system. These coordinate values are directly measured by the laser scanning device and obtained through coordinate transformation. For example, for the th feature point, its laser scanning coordinates are .

[0020] is the three-dimensional coordinates of the corresponding th point in the image measurement sequence, indicating that in the "list of aligned joint feature point pairs", the eastward coordinate, northward coordinate, and elevation value of the homologous point from the unmanned aerial vehicle image sequence and solved by photogrammetry methods (such as SfM or binocular vision) in the th pair of joint feature points in the WGS-84 coordinate system. These coordinates are the three-dimensional point coordinates obtained after aerial triangulation and bundle adjustment of the unmanned aerial vehicle aerial images. For example, for the th feature point, its image measurement coordinates are .

[0021] Calculation process: It is set that 3 pairs of joint feature points are selected from the "list of aligned joint feature point pairs" for calculation, that is .

[0022] The first pair of feature points ( ): Laser scanning coordinates ; Image measurement coordinates ; The spatial distance of the first pair of points

[0023] ; ; ; The second pair of feature points ( ): Laser scanning coordinates ; Image measurement coordinates ; Spatial distance between the second pair of points ; ; ; ; The third pair of feature points ( ): Laser scanning coordinates ; Image measurement coordinates ; Spatial distance between the third pair of points ; ; ; ; Calculate the registration factor of the rail profile curve : ; ; ; ; The result shows that: in this example, the average three-dimensional spatial deviation between the laser scanning data and the UAV image measurement data at the selected 3 pairs of combined feature points is about 0.0306 m (or 30.6 mm), and this value reflects the overall geometric accuracy of the current multi-source data fusion.

[0024] According to the rail profile curve registration factor calculated in the previous step and the three-dimensional coordinates of each pair of feature points in the "list of aligned combined feature point pairs", first calculate the three-dimensional Euclidean distance between each pair of feature points, that is, the three-dimensional offset value, and its calculation method is , then compare the three-dimensional offset value calculated for each point pair with a preset screening threshold, and the screening threshold here is set to the rail profile curve registration factor a specific multiple, such as 1.5 times, which is the screening threshold , if in the example , then , if the three-dimensional offset value of a point pair is greater than this , then this point pair is determined as a gross error data point and removed from the list, or marked as having a low weight to reduce its impact on subsequent processing. Next, for the valid feature point pairs retained after screening, recalculate the remapping vector in the global WGS-84 coordinate system and perform coordinate replacement. This step aims to unify and optimize the coordinates of the feature points. Select a data source with higher precision (such as lidar scan data) as the benchmark. If the lidar point coordinate is and the image point coordinate is , and this point pair is not screened out, then the lidar point coordinate can be used as the final fused coordinate of this feature point, that is, the remapped coordinate is , or calculate the weighted average of the two, and the weights are determined according to their respective prior precisions. For example, the precision of lidar point cloud is usually better than that of UAV photogrammetry. Therefore, a higher weight can be given to the lidar data, such as 0.7, and the weight of UAV data is 0.3. Then the remapped coordinate is . At the same time, perform coordinate fitting and correction on the geometric parameters (mainly referring to its three-dimensional coordinates) of the center point of the sleeper cross-section. Project the coordinates of a series of sleeper center points distributed along the track onto the track center line fitted through these points, or perform plane fitting on the sleeper point cloud in a local area to correct its elevation and plane position, and eliminate isolated measurement errors. For example, the mathematical expression of a section of the track center line can be fitted by the least squares method, and then each sleeper center point is projected onto this fitted line to obtain the corrected sleeper center point coordinates. Finally, integrate the rail profile point coordinates after remapping processing with the geometric parameters of the sleeper cross-section center point after coordinate fitting and correction to generate an integrated structure attribute set.

[0025] The steps to obtain the geometric deviation metric element are as follows: Based on the preset standard geometric shapes of the rail and the sleeper in the digital twin model, call the rail edge profile point coordinates and the geometric parameters of the sleeper cross-section center point in the integrated structure attribute set, match the corresponding rail component reference points and the sleeper component center points in the digital twin model, and form a space point pair with a one-to-one correspondence between the model and the measured points; Based on the coordinates of each pair of rail component reference points and the measured rail edge profile points in the space point pair, calculate the coordinate differences of this point pair along the X, Y, and Z axes in the three-dimensional space one by one, synthesize a complete three-dimensional space displacement vector and record the vector length, and at the same time calculate the average space displacement vector length of the point pair set to generate a rail space displacement vector set; Based on the geometric parameters of the center points of each pair of sleeper components and the center points of the measured sleeper cross-sections in the spatial point pairs, the Euler angle difference values of the rotational attitude changes between the two points are extracted, and the attitude angle differences in the pitch angle, yaw angle, and roll angle are respectively recorded to form a set of sleeper attitude angle differences. Then, the set of rail spatial displacement vectors is combined with the set of sleeper attitude angle differences to obtain the geometric deviation metric elements.

[0026] Specifically, based on the standard geometric shapes of the rail and the sleeper preset in the digital twin model, first, the obtained integrated structure attribute set is called. This attribute set contains the three-dimensional coordinates of the rail edge contour points and the geometric position parameters of the center points of the sleeper cross-sections after multi-source data fusion processing. Subsequently, a matching operation is performed. The predefined corresponding rail component reference points in the digital twin model (these reference points are arranged on the theoretical center line or the rail top surface of the rail according to the design drawings during the construction stage of the digital twin model. For example, a three-dimensional coordinate point is defined as a reference point every 0.25 meters along the rail center line) are associated with the measured rail edge contour point coordinates extracted from the integrated structure attribute set. At the same time, the center points of the sleeper components in the digital twin model (i.e., the preset three-dimensional coordinates of the sleeper geometric center and the standard installation attitude parameters, such as the Euler angles in the ideal state) are matched with the geometric parameters of the measured sleeper cross-section center points in the integrated structure attribute set (including the measured three-dimensional coordinates and the actual attitude parameters obtained through local point cloud analysis). For each measured rail edge point, the system searches for the reference point with the closest spatial distance on the corresponding rail theoretical profile in the digital twin model. This search is limited within a dynamically adjusted search window. The initial search radius is set to 0.1 meters, for example. The setting of this radius refers to the deviation distribution range of 95% in the historical geometric state detection data of the same type of track. If the deviation of the matching point pair is greater than this radius, the search window may be enlarged or marked as a potential gross error point. For the sleeper, a preliminary match is made based on the three-dimensional Euclidean distance between the measured sleeper center point and the preset sleeper center point in the model. The search radius is set to 0.15 meters, for example. This value is determined based on the upper limit of the allowable deviation of the sleeper position (for example, the allowable lateral deviation is ±10 mm, the allowable elevation deviation is ±5 mm, and comprehensive considerations such as planar distortion) and combined with a certain safety margin (for example, increasing by 50%). To ensure the uniqueness and accuracy of the match, the system also uses local geometric feature information such as the normal vector and curvature of the point cloud, or the sequence relationship of the point positions (such as the continuity along the rail direction) to assist in the judgment. When the distances from a measured point to multiple model reference points are all within the threshold, the one with the most similar geometric features or the most consistent sequence is selected as the final matching pair to form a one-to-one spatial point pair between the model and the measured points.

[0027] Based on the one-to-one corresponding spatial point pairs obtained in the previous step, which are composed of the reference points of the rail components in the digital twin model and the coordinates of the measured rail edge contour points, the system processes each pair of such spatial point pairs. Specifically, if the coordinates of the th reference point of the rail component in the model are , and the coordinates of the corresponding measured rail edge contour point are , then first, calculate the coordinate differences of this pair of points along the three orthogonal coordinate axes X, Y, and Z in the three-dimensional space one by one, that is , , and . These three difference components together form the three-dimensional space displacement vector from the model reference point to the measured point. This vector intuitively represents the spatial deviation of the measured point from its ideal design position. Subsequently, calculate and record the length (i.e., the modulus) of this three-dimensional space displacement vector . This length quantifies the absolute spatial deviation magnitude between this pair of points. After completing the above calculations for all spatial point pairs, the lengths of all calculated three-dimensional space displacement vectors will also be statistically analyzed to calculate the average value of these length values, that is, the average three-dimensional space displacement vector length of the point pair set, where is the total number of spatial point pairs. This average length value can be used as a macroscopic indicator to evaluate the overall geometric compliance of the rail. Finally, organize all the calculated individual three-dimensional space displacement vectors and their corresponding lengths , together with the calculated average three-dimensional space displacement vector length , to generate a set of rail three-dimensional space displacement vectors.

[0028] Based on the one-to-one corresponding spatial point pairs between the model and the measured points generated in the previous step, especially for the part of the sleeper components, that is, each pair consists of the center point of the sleeper component in the digital twin model (including its ideal three-dimensional coordinates and a preset standard rotation attitude, which can be represented by a standard rotation matrix or a set of reference Euler angles such as ) and the geometric parameters of the center point of the measured sleeper cross-section obtained from the integrated structural attribute set (including the measured three-dimensional coordinates and the actual rotation attitude obtained by analyzing the sleeper point cloud data or directly measuring with sensors, which can be represented by an actual rotation matrix or a set of measured Euler angles ), the system will extract and quantify the change in the rotation attitude between the two. First, calculate the relative rotation transformation from the model attitude to the measured attitude. If the attitude is represented by a rotation matrix, the differential rotation matrix can be calculated as (or , because for a rotation matrix, its inverse is equal to its transpose), from this differential rotation matrix it is possible to resolve the Euler angles representing the attitude difference, namely the pitch angle difference ( ), the yaw angle difference ( ), and the roll angle difference ( ). This resolution process uses the standard conversion formula from a rotation matrix to Euler angles and requires attention to handling special cases such as gimbal lock to ensure the uniqueness and correctness of the angles. The attitude angle difference values on these calculated pitch, yaw, and roll angles are recorded one by one to form a set of sleeper attitude angle differences. Finally, this set of sleeper attitude angle differences is combined with the set of rail spatial displacement vectors (including the displacement vectors of each reference point on the rail and their length information) generated in the previous step. This combination integrates two different types of deviation information (the linear displacement deviation of the rail and the angular attitude deviation of the sleeper) into a unified data structure, such as forming a list containing the geometric deviations of each sleeper and its associated rail segment, to obtain the geometric deviation metric element.

[0029] The steps for obtaining the calibrated geometric representation are as follows: Based on the geometric deviation metric element, extract each spatial displacement vector in the set of rail spatial displacement vectors, and respectively resolve the displacement components of this spatial displacement vector on the three coordinate axes of X, Y, and Z. For each rail node coordinate in the digital twin model, add the resolved displacement component values of each axis to the original coordinate value of the node to form an adjusted sequence of rail node coordinates; Based on the adjusted sequence of rail node coordinates, call the set of sleeper attitude angle differences in the geometric deviation metric element, and resolve the pitch angle difference, yaw angle difference, and roll angle difference of each sleeper in the set one by one, and recalculate the rotation matrix parameters of each sleeper component to generate a corrected sequence of sleeper rotation matrices; Based on the adjusted sequence of rail node coordinates and the corrected sequence of sleeper rotation matrices, apply the sequence of rail node coordinates and the sequence of sleeper rotation matrices to the original geometric configuration of the rail and sleeper in the digital twin model one by one, and perform a three-dimensional shape mapping update again to form the calibrated geometric representation.

[0030] Specifically, based on the geometric deviation metric element obtained in the previous step, first extract the set of rail spatial displacement vectors from it. For each spatial displacement vector in this set, such as the th spatial displacement vector , the system will respectively resolve its displacement components on the three coordinate axes of X, Y, and Z in the global coordinate system, namely , and , these components directly represent the deviations of the measured points on the rail with respect to the corresponding reference points in the digital twin model. Then, traverse all the preset rail nodes in the digital twin model. For each rail node, for example, its original coordinates are , if there is a clear corresponding relationship between this node and a certain spatial displacement vector (whose components are ) (this corresponding relationship is established when calculating the geometric deviation metric element, that is, the matching of the model reference point and the measured point), then directly add the numerical values of the displacement components on each axis parsed to the original coordinate values of this node, that is, the adjusted node coordinates . If there are multiple measured point displacement vectors near a model node, the comprehensive displacement correction amount of this node may be calculated by means of weighted average (such as weighted by the reciprocal of the distance), or according to the predefined mapping rule (for example, interpolate all the measured point displacement vectors on a section of the rail to the discrete model nodes of this section of the rail), update the coordinates of all relevant rail nodes. By performing this coordinate superposition operation on all the rail nodes in the digital twin model that have a corresponding relationship with the measured data, an adjusted sequence of rail node coordinates is formed.

[0031] Based on the adjusted sequence of rail node coordinates obtained in the previous step, the system then calls the set of sleeper attitude angle differences in the geometric deviation metric element. This set records the pitch angle difference, yaw angle difference, and roll angle difference of each measured sleeper with respect to its preset standard attitude in the digital twin model. The system will process the attitude angle difference data of each sleeper in the set one by one. For example, for the th sleeper, the recorded pitch angle difference is , the yaw angle difference is , and the roll angle difference is . At the same time, obtain the original rotation matrix of this sleeper in the digital twin model, denoted as (or its original Euler angles ). Then, use these angle difference values to recalculate the rotation matrix parameters of this sleeper component. If both the original attitude and the differences are given in Euler angles, the difference Euler angles can be converted into a difference rotation matrix , and then update the original rotation matrix through matrix multiplication: (Note the convention of the rotation order. Usually, the intrinsic rotation is applied first, then the deviation rotation, or directly add the difference angles to the original angles and then convert them into a new rotation matrix. However, directly adding Euler angles may introduce problems when there are large angle rotations or close to gimbal lock, so the operation of converting to a rotation matrix is more reliable). For example, the original Euler angles can be converted into the original rotation matrix , and then the difference Euler angles Convert to difference rotation matrix , and finally calculate the adjusted rotation matrix , the rotation matrix recalculation process is performed on all sleeper components with attitude angle difference records to generate a corrected sleeper rotation matrix sequence.

[0032] Based on the adjusted rail node coordinate sequence and the corrected sleeper rotation matrix sequence obtained in the first two steps, the system applies these two sets of updated geometric parameters one by one to the original geometric configuration of rails and sleepers in the digital twin model. In the specific operation, for the rail part, the coordinates of each node of the rail centerline or contour line originally defined by a series of nodes are replaced by the corresponding values ​​in the adjusted rail node coordinate sequence, which directly leads to the update of the rail geometry in the digital twin model, making it closer to the actual measured state. For the sleeper part, the spatial posture of each sleeper component in the model is determined by its rotation matrix. Now its original rotation matrix is ​​replaced with the new rotation matrix of the corresponding sleeper in the corrected sleeper rotation matrix sequence. The center position of the sleeper may also need to be adjusted in conjunction with the displacement of its associated rail node, or the coordinates of the sleeper center point corrected by the integrated structural attributes can be directly used. After updating the coordinates of all relevant rail nodes and the rotation matrices of all relevant sleeper components, the system will trigger a re-rendering or geometric reconstruction process of the 3D model. In this process, the model rendering engine or geometric modeling kernel will regenerate the 3D entity or surface of the rail according to the new node coordinates, and reposition and orient the 3D entity or surface of each sleeper according to the new rotation matrix and position parameters. This re-update of the 3D morphological mapping ensures that the digital twin model is consistent with the latest measurement status of the actual track structure at the geometric level, forming a calibrated geometric representation.

[0033] The steps to obtain the connection change instruction set are: Retrieve the coordinates of the rail nodes and the rotation matrix of the sleeper components in the calibration geometric representation, and identify the adjacency information and connection type identification in the original connection relationship between each node and each sleeper one by one. Combined with the text content of the maintenance record during the sleeper-rail maintenance period, extract all record sentences with the keywords "replacement", "removal" and "addition", and parse the associated sleeper numbers and fastener numbers to form an initial list of structural connection changes; Based on the initial list of structural connection changes, according to the time period and spatial position interval corresponding to each sleeper number, the parsed sleeper number and fastener number are compared with the current connection status in the calibration geometric representation one by one. If there is a sleeper number in the list but not in the calibration geometric representation, it is determined to be a newly added sleeper connection. If there is a fastener number in the calibration geometric representation but not in the list, it is determined to be a fastener removal, and a list of component connection relationship changes is generated; According to the component connection relationship change list, construct a standardized instruction text format for all connection relationships marked as newly added or removed, and sequentially supplement the instruction type, target number, operation node index, and scope of influence identifier to generate a connection change instruction set.

[0034] Specifically, retrieve the calibrated geometric representation generated in the previous step. This representation contains the updated rail node coordinates and sleeper component rotation matrices. Based on this, first analyze the original connection relationships between each rail node and each sleeper in the model, and item by item identify the adjacent information of each connection. For example, clearly indicate which fastener position of which sleeper a certain rail node is directly connected to, and the connection type identifier, such as "elastic fastener connection", "rigid connection", or other specific types of constraints. These original connection information is usually preset according to the design specifications during the initialization of the digital twin model. Next, the system combines the text content of the maintenance records within the sleeper-rail maintenance cycle imported from an external database or file. This text content records historical maintenance activities, such as what operations were carried out on a certain section on a certain date. Using keyword matching and named entity recognition methods, extract all statements containing the preset keywords from these maintenance record texts. These keywords are set as "replacement", "removal", and "new addition". For example, when detecting a record like "On May 10, 2023, 3 sleepers (numbers S001, S002, S003) were replaced at K10+500, and 2 sets of fasteners (numbers F005, F006 on sleeper S001) were newly added", the keywords "replacement" and "new addition" will be recognized, and further analyze the specific sleeper numbers (such as S001, S002, S003) and fastener numbers (such as F005, F006) associated with these operations, as well as the location (K10+500) and time where the operation occurred. By performing such automated extraction and analysis on all maintenance record texts, summarize all the identified structural change information to form an initial list of structural connection changes.

[0035] Based on the initial list of structural connection changes generated in the previous step, the system will meticulously compare each record in the list with the component connection status of the current record in the calibrated geometric representation one by one, according to the timestamp and spatial location information (such as kilometer post, line name, etc.) corresponding to the sleeper number associated with each record. This comparison process involves aligning the time information in the maintenance records with the specific time snapshot represented by the calibrated geometric representation. For example, if the calibrated geometric representation represents the state on June 1, 2023, then in the initial list of structural connection changes, all relevant maintenance records earlier than this date should be considered. For spatial location, the system will locate the corresponding sleeper or rail section in the calibrated geometric representation based on the sleeper number or section information provided in the maintenance records. Then, for each sleeper number mentioned in the maintenance records, it will check whether this number exists in the current sleeper list of the calibrated geometric representation. If a certain sleeper number appears in the initial list of structural connection changes (usually due to "replacement" or "new addition" records), but the corresponding sleeper entity cannot be found in the calibrated geometric representation by its unique identifier (such as the sleeper number), it is determined that the sleeper was newly added to the track structure after the time of the maintenance record and before the time of the current representation, and it is recorded as a new sleeper connection. Conversely, for fasteners, it will check the current fastener connection relationships in the calibrated geometric representation. If a certain fastener number (for example, uniquely determined by its installation position and type on the sleeper) exists in the connection records of the calibrated geometric representation, but the initial list of structural connection changes indicates that the fastener has been "removed" or its associated sleeper has been "replaced" (replacing a sleeper usually means removing the old fasteners on it), it is determined that the fastener has been removed and recorded as a fastener removal. Through this two-way comparison, it is possible to identify the component addition and subtraction changes caused by maintenance activities, and organize and record these changes to generate a list of component connection relationship changes.

[0036] Based on the list of changes in component connection relationships generated in the previous step, which details all the sleeper and fastener connection relationships determined to be newly added or removed, the system then processes each change record in the list. The goal is to construct a standardized instruction text format so that the subsequent topology evolution module can accurately understand and execute these changes. For each connection relationship marked as "newly added" (such as adding a new sleeper or a set of fasteners), the system generates an instruction. This instruction first includes the instruction type, clearly identified as "ADD_CONNECTION", and then supplements the target number, which is the unique identifier of the newly added component (such as the number S101 of the new sleeper, or the position number C01-L of the new fastener on its affiliated sleeper). Next is the operation node index, which refers to the index number or ID of the rail node or sleeper component associated with this newly added connection in the digital twin model. Finally, there is the scope of influence identifier, which is used to describe the local area or adjacent components affected by this newly added connection. For example, adding a new sleeper will affect the constraints of the rail segments on both sides, and adding a fastener will change the connection characteristics between a specific rail and a specific sleeper. For each connection relationship marked as "removed" (such as removing an old sleeper or a set of damaged fasteners), the system also generates an instruction. The instruction type identifier is "REMOVE_CONNECTION", the target number is the identifier of the removed component, the operation node index indicates the model node affected by this removal operation, and the scope of influence identifier explains which connections no longer exist or have changed characteristics after the removal. By performing this process of constructing the standardized instruction text for all entries in the list of changes in component connection relationships and supplementing the complete instruction elements in sequence, a connection change instruction set with a clear structure and clear content is finally formed.

[0037] The steps to obtain the evolved topological network diagram are as follows: Based on the connection change instruction set, parse the target number, operation node index, and scope of influence identifier in each connection change instruction in sequence, call the rail node coordinates and sleeper component index in the calibrated geometric representation body, and determine whether the target number corresponds to the existing node set. If it is a new addition instruction, add this number to the node set and construct an initial connection entry. If it is a removal instruction, delete all connection entries associated with this number to obtain the structure connection reconstruction node table; According to the structure connection reconstruction node table, scan the connection status between each rail node coordinate and the affiliated sleeper component, extract all node pairs with physical connection relationships, and calculate the connection orientation attribute according to the rotation matrix state of the sleeper component to form a constraint relationship reconstruction entry table; Based on the constraint relationship reconstruction entry table, establish an adjacency list structure with the node number as the main index one by one, embed the connection status, connection direction, and connection label as edge attributes into the corresponding adjacency entries, and simultaneously construct an edge index table to generate the evolved topological network diagram.

[0038] Specifically, based on the connection change instruction set obtained in the previous step, the system will sequentially read and parse each connection change instruction. For each instruction, it extracts the target number contained therein (e.g., the unique ID of the sleeper or fastener to be added or removed), the operation node index (indicating the internal index or ID of the rail node or sleeper component directly associated with this change in the digital twin model), and the scope of influence identifier (describing the local topological structure affected by this change, such as a list of adjacent component IDs or a spatial region description). While parsing the instruction, the system will call the rail node coordinate data and sleeper component index information in the calibrated geometric representation body, which constitute the geometric and identification basis of all structural components in the current digital twin model. Then, according to the target number in the instruction, the system needs to determine whether the component represented by this number already exists in the existing node set of the calibrated geometric representation body (i.e., the list of all rail nodes and sleeper components in the current model). If the instruction type is an addition instruction (e.g., the instruction text starts with "ADD_CONNECTION"), and the target number is confirmed not to be in the existing node set after querying, the system will add this new target number to the node set and construct one or more initial connection entries according to the operation node index and scope of influence identifier in the instruction. These initial connection entries define the basic connection relationships between the new component and its surrounding components. For example, when adding a sleeper S101, its operation node index may point to the rail nodes G50 and G51 on both sides of it, and the scope of influence identifier may indicate that it is connected through a standard fastener type. If the instruction type is a removal instruction (e.g., the instruction text starts with "REMOVE_CONNECTION"), and the target number is confirmed to exist in the existing node set, the system will delete this target number from the node set and simultaneously delete all connection entries associated with this number. This means that all connections pointing to the removed component and connections originating from this component will be cleared. By performing the above judgments and operations on all instructions in the connection change instruction set, the nodes and their basic connection information in the model are dynamically updated to obtain the structure connection reconstruction node table.

[0039] The structure connection reconstruction node table updated and obtained according to the previous step, which reflects the current existing rail nodes and sleeper components in the model and their preliminary connection intentions after component addition and subtraction due to maintenance activities. Next, the system will conduct a detailed scan and analysis of the coordinate information of each rail node in the model and the connection status between the rail node and the corresponding sleeper component it may connect to. The purpose of this scan is to identify and extract all the actual existing physical connection relationships to form specific node pairs. For example, the system will check whether there is a sleeper component (also defined by its index in the table and the associated rotation matrix) near each rail node (defined by its coordinates in the structure connection reconstruction node table), and based on a preset connection judgment criterion (for example, the spatial distance between the rail node and the fastener installation position on the sleeper is less than a specific connection tolerance threshold, which is set to 5 mm. This value is comprehensively considered based on the precision requirements of fastener installation and the actual measurement error. For example, if the allowable installation clearance of the fastener is 3 mm and the measurement error is 2 mm, then the tolerance threshold is 5 mm), to determine whether there is a physical connection between the two. Once it is determined that there is a physical connection between a pair of rail nodes and a sleeper component (or more specifically, a certain fastener position on the sleeper), the system will record this pair of nodes and further calculate the orientation attribute of this connection based on the rotation matrix state of the sleeper component (the rotation matrix has been updated in the calibrated geometric representation and reflects the actual spatial attitude of the sleeper). The orientation attribute describes the direction and relative attitude of the connection. For example, the representation of the vector pointing from the rail node to the fastener contact point in the local coordinate system of the sleeper can be calculated, or the contact angle of the rail relative to the sleeper can be described, etc. By conducting such a connection status scan and orientation attribute calculation for all rail nodes and sleeper components, the system summarizes all the identified effective physical connections and their orientation attributes to form a constraint relationship reconstruction entry table.

[0040] Reconstruct the entry table based on the constraint relationships generated in the previous step. This table contains all the actual physical connection information that has been verified and had its attributes calculated. The system now begins to construct the final evolving topological network diagram. Specifically, first, for each node in the network diagram (representing a rail node or a sleeper component), an adjacency list structure is established with its unique node number (derived from the structure connection reconstruction node table) as the main index. An adjacency list is a commonly used graph representation method where each node is associated with a list that stores the numbers of all other nodes directly connected to that node. During the construction of the adjacency list, the system will embed the connection status extracted from the constraint relationship reconstruction entry table (such as "normal connection", "loose", "damage warning", etc., and this status information may come from further parsing of maintenance records or sensor data), as well as the previously calculated connection direction (i.e., part of the connection orientation attribute, such as a unit vector or a specific angle representing the connection direction) and connection label (e.g., "Connection of the left rail to the No. 1 fastener of sleeper A", "Elastic fastener type II") as the attributes of the edge into the corresponding adjacency entries. This means that when recording node B in the adjacency list of node A, these detailed attribute information of the connection edge between A and B will also be attached. To improve query efficiency and facilitate subsequent analysis, while constructing the adjacency list, the system will also synchronously construct an edge index table that stores information about all the edges in the network. Each edge is uniquely identified by the pair of numbers of the two nodes it connects and is associated with all its attributes. By processing each entry in the constraint relationship reconstruction entry table one by one and filling the adjacency list and edge index table in the above manner, the evolving topological network diagram reflecting the latest topological connection relationship of the current rail-sleeper system is finally constructed completely.

[0041] The steps to obtain the hybrid fidelity model configuration are as follows: Based on the node structure relationships in the evolving topological network diagram, extract the set of connection edges between all node pairs, and identify the node numbers and spatial distribution positions associated with each connection edge one by one. Match the structural fatigue monitoring parameter table given in the rail-sleeper maintenance cycle simulation analysis requirements, and screen out the set of regional edges where the connection edges fall into the fatigue risk identification interval to generate a stress region boundary index list; Based on the stress region boundary index list, calculate the track element model adaptation factor for each stress region. The calculation formula is: ; Where, is the track element model adaptation factor for the stress region, is the total number of nodes associated with the connection edges within the stress region, is the th measured vertical dynamic load value of the node during the current monitoring period, is the The cumulative stress value of a node within the current maintenance cycle is the theoretical stress reference value defined for the equivalent acting area of the structural contact region is the characteristic length of the rail micro-segment represented by the th node, defined as where represents the physical length of the rail segment corresponding to the th node truly mapped in the topology diagram, and represents the minimum characteristic length limit value set in the simulation configuration; Based on the track unit model adaptation factor corresponding to each stress region, compare the track unit model adaptation factor with the adaptation value reference threshold item by item. If the track unit model adaptation factor corresponding to the region is greater than or equal to the adaptation value reference threshold, then designate the region as a track unit model with a complete parameter configuration. Otherwise, designate it as a track unit model with a parameter compression configuration, and integrate the two types of track unit models into a hybrid fidelity model configuration of the overall track system.

[0042] Specifically, based on the nodes (representing rail nodes or sleeper components) and their connection relationships contained in the evolutionary topology network diagram obtained in the previous step, the network diagram is first traversed to extract all the edges representing direct physical connections between the nodes, forming a set of connection edges. For each connection edge in this set, the system will successively identify the unique numbers of the two end nodes associated with it and the precise distribution positions of these nodes in three-dimensional space. These spatial position information directly comes from the rail node coordinates and sleeper component geometric parameters in the calibrated geometric representation body. Subsequently, the system will match the structural fatigue monitoring parameter table predefined in the simulation analysis requirements for the rail-sleeper maintenance cycle. This parameter table is compiled based on the common disease locations summarized from historical maintenance data (e.g., welded joints, wear areas of the outer rail of small-radius curves), listing the fatigue sensitivity indicators and their corresponding risk level thresholds of different structural components under different working conditions. For example, the upper limit of the allowable crack growth rate in the rail welded joint area is 0.01 mm per day, or the warning value of the rolling contact fatigue index for rail tread contact fatigue is 0.8. The entries in these parameter tables define specific fatigue risk identification conditions. The current state parameters of each connection edge and its associated nodes (such as cumulative stress, number of passes, known defects, etc., which are obtained from the historical database or real-time monitoring data of the digital twin model) are matched with the risk identification conditions in the structural fatigue monitoring parameter table. When the relevant parameter values in the area where a certain connection edge is located reach the preset fatigue risk identification interval, for example, the cumulative total weight of a certain section of rail exceeds 80% of the design life (such as the design total weight is 6 billion ton-kilometers and the current has reached 4.8 billion ton-kilometers), or the stress concentration factor of its associated node exceeds 1.5 (this factor is pre-calculated by finite element analysis or estimated according to geometric irregularities), then this connection edge is determined to fall into the fatigue risk identification interval. All the connection edges so determined are collected to form a set of regional edges. Finally, based on these selected sets of regional edges, the system summarizes and removes duplicates of the end nodes of these edges and groups them according to their spatial order or logic on the track line to generate a stress area boundary index list.

[0043] Formula: , the advantage of the formula is that by comprehensively considering the instantaneous dynamic load and cumulative stress deviation of the node and combining its geometric characteristic length, an adaptation factor of the track element model is calculated for each potential stress area , this factor can quantify the necessity and urgency of high-fidelity simulation in this area. In the formula, the dynamic load and the cumulative stress deviation term take the square root of the sum of squares, similar to calculating the modulus of a two-dimensional vector, integrating the contributions of short-term impact and long-term damage, and multiplying by the equivalent action area to convert the stress deviation into an equivalent force, making it consistent with Dimensionally comparable, divided by the characteristic length The range of stress influence or element size effect is considered, making the factor sensitive to both local high stresses and widely distributed sub-high stresses. The final average value enables It to represent the comprehensive state of the entire stress area, providing a basis for subsequent adaptive selection of simulation models with different fidelities.

[0044] is the total number of nodes associated with the connecting edges within the stress area. This parameter is obtained by counting the unique number of nodes included in the "stress area boundary index list" generated in the previous step. These nodes are the structural key points that make up a specific stress area. For example, if a stress area contains 3 connecting edges, and these 3 edges jointly connect 4 different rail nodes or sleeper components, then .

[0045] is the th measured vertical dynamic load value of the node during the current monitoring period. This data is collected in real time by sensors deployed on the track structure (such as strain gauges pasted on the rail foot or web, fiber Bragg grating sensors, or accelerometers installed on the sleepers). When a train passes, the sensors record the dynamic response data, and the maximum vertical dynamic load or an equivalent dynamic force amplitude borne by each monitored node within a specific time window (e.g., during one train passage) is extracted. For example, by analyzing the data of a strain sensor at a certain node during the most recent train passage, the peak strain is extracted and multiplied by the corresponding force-strain conversion coefficient to obtain .

[0046] is the th cumulative stress value of the node during the current maintenance cycle. This value is calculated by combining the historical load spectrum, the S-N curve (stress-life curve) of the material, and the linear cumulative damage theory (such as Miner's rule). Since the last major repair or component replacement, the number of various load cycles experienced by the node and their corresponding stress amplitudes are recorded, and the current cumulative damage degree or equivalent cumulative stress is estimated accordingly. Example of the calculation formula: , where is the actual number of cycles at the stress level , is the allowable number of cycles at this stress level (obtained from the S-N curve), is a reference stress value, or directly track the equivalent damage degree. For example, a node has endured loads equivalent to 60% of the design fatigue life during the current maintenance cycle, and its design fatigue limit stress is 250 MPa. Then its current cumulative stress state can be estimated .

[0047] For the The theoretical stress reference value of a node defined in the current structural standard represents the expected stress level of this type of node under ideal conditions and under standard design loads, or serves as a stress threshold for fatigue assessment (stress below this value is not included in fatigue accumulation). For example, for a 60kg / m rail, under the action of a train with a standard axle weight of 25 tons, the theoretical calculated stress reference value at the lower jaw of the rail head is set to .

[0048] is the equivalent effective area of ​​the structural contact area. This parameter characterizes the effective contact range when the load is transferred between components. For wheel-rail contact, the area of ​​the wheel-rail contact patch can be calculated using the Hertz contact theory. For the contact between the fastener and the rail or sleeper, the effective bearing area under the rated preload is estimated based on the fastener design drawings and the material elastic modulus. This value is usually determined based on the specific contact pair type and geometry. For example, for a typical elastic cushion between the rail and the sleeper, its equivalent effective area under the standard buckling pressure is (i.e. 100cm ).

[0049] For the The characteristic length of the rail micro-segment represented by the node is defined as ,in Indicates The physical length of the rail segment that each node actually corresponds to in the evolutionary topology network diagram. This length is determined according to the discretization degree of the rail in the digital twin model. For example, if a node is set every 0.5 meters along the length of the rail, then , It is a minimum characteristic length limit value set in the simulation configuration to avoid numerical calculation problems caused by excessive discretization of the model. For example, according to experience or simulation stability requirements, , then if , .

[0050] Calculation process: With a The stress area of ​​a node is calculated as an example.

[0051] Node 1( ): ; ; ; (For example, the same for all nodes); ; Node 2( ): ; ; (The reference values at different positions may be different); ; ; First, calculate the contribution terms of each node : For ease of calculation, convert MPa to kN / m (1 MPa = 1000 kN / m ).

[0052] .

[0053] For node 1: ; ; ; ; ; For node 2: ; ; ; ; ; Calculate : ; ; ; ; The result shows that the track element model adaptation factor of this stress region is approximately , and this value comprehensively reflects the severity of the dynamic load effects and cumulative stress deviations of the nodes in this region.

[0054] Based on the track element model adaptation factors corresponding to each stress region calculated in the previous step , the system will compare these item by item values with a preset adaptive value benchmark threshold, which is comprehensively set based on historical simulation data, maintenance experience, and the trade - off between simulation accuracy and computational cost. The setting process is as follows: collect the calculated values of 100 different track sections in history and their subsequent actual structural states (determined as "high - risk requires detailed analysis" or "low - risk can be simplified for analysis" through inspections or monitoring), and record the computational time and prediction accuracy of simulation analysis using the full - parameter configuration model and the parameter - compressed configuration model. Set an evaluation function , where is the candidate adaptive value benchmark threshold, represents the average accuracy improvement rate brought by using the high - fidelity model compared to the low - fidelity model in the region of represents the average computational time increase ratio due to this, and the weights and reflect the preference for accuracy and efficiency. For example, set , by testing a series of values (such as from the 50th percentile to the 95th percentile of the historical distribution, with a step size of the 5th percentile), select the value that maximizes as the final adaptive value benchmark threshold. For example, after the above analysis, determine the adaptive value benchmark threshold to be . During the comparison process, if the track element model adaptation factor calculated for a certain stress region is greater than or equal to (for example, the calculated in the previous example), then ), the system determines that this region requires high - precision simulation and designates it as a track element model with a full - parameter configuration, which will include detailed geometric structures, accurate material constitutive relations, and complex contact and boundary conditions. Conversely, if the corresponding to the region is less than , it is considered that the structural risk of this region is relatively low, or its behavior has a relatively small impact on the overall system, and a track element model with a parameter - compressed configuration can be used. This model may use simplified geometries, linear elastic materials, or equivalent macroscopic mechanical parameters to reduce computational complexity. After performing this judgment and designation operation on all stress regions, the system organically combines these two different - fidelity track element models (the full - parameter configuration model and the parameter - compressed configuration model) according to their spatial positions and connection relationships in the actual track, and integrates them into a hybrid - fidelity model configuration of the overall track system.

[0055] The steps to obtain the adaptive simulation output results are as follows: Based on each region delimited by the hybrid fidelity model configuration, the track unit models with complete parameter configuration or parameter compression configuration are called region by region. The current load conditions and material parameters of each track unit model within the region are respectively loaded, and dynamic stress analysis and calculation are carried out to obtain the stress data sequence of each region's nodes in real time, and a real-time data set of regional node stress is generated; According to the real-time data set of regional node stress, calculate the difference in stress numerical values within consecutive time steps in each region and divide it by the corresponding time interval to obtain the stress gradient change rate of each region. Compare the calculated stress gradient change rate with the preset stress gradient change rate threshold region by region to identify the abnormal regions where the change rate exceeds the preset limit, and form a list of identified abnormal stress regions; Based on the list of identified abnormal stress regions, for each abnormal region in the list, upgrade the track unit model originally configured with parameter compression configuration to a complete parameter configuration, refine the parameters of the track unit model originally configured with a complete parameter configuration, update the stress data again, and generate an adaptive simulation output result.

[0056] Specifically, based on the hybrid fidelity model configuration obtained in the previous step, which specifies which type of track unit model (complete parameter configuration or parameter compression configuration) should be used in each spatial region of the track system, the system will call the corresponding track unit models region by region according to this configuration. For the regions designated with a complete parameter configuration, the system will instantiate a finite element model or a multibody dynamics model that includes detailed geometry (such as precise rail and sleeper cross-sections, refined fastener connections), non-linear material constitutive relations (such as considering the elastic-plastic and fatigue characteristics of steel), complex contact algorithms (such as wheel-rail Hertz contact or more refined non-Hertz contact models), and precise boundary conditions. For the regions designated with a parameter compression configuration, a simplified beam element model, an equivalent spring-damper system, or a macroscopic mechanical model based on empirical formulas may be called. After calling the model, the system will load the current load conditions for each track unit model within the region, and these load conditions include the time series of dynamic wheel loads generated by train operation (obtained from the vehicle dynamics simulation module or actual monitoring data, for example, providing the three-directional forces of the wheel at each time step ), additional stresses caused by temperature changes (calculated based on the ambient temperature and rail restraint conditions), and additional excitations caused by track irregularities, etc. At the same time, the latest material parameters are loaded, and these parameters may be adjusted according to the service life, cumulative damage, or environmental factors of the components (for example, the elastic modulus of the material changes with temperature, or fatigue damage leads to a reduction in material strength). After the loading is completed, the system performs independent or coupled dynamic stress analysis and calculation on the models of each region, using numerical integration methods (such as Newmark- Solve the dynamic response equation of the structure by a method (such as explicit / implicit time integration) to obtain in real-time the stress (such as Mises equivalent stress, principal stress) data series of each key node (such as specific monitoring points on the rail, sleeper, and fastener) in the simulation time history within this region. These series record stress values at a fixed time step (for example, 0.001 seconds), and finally integrate the node stress data series of all regions to generate a real-time data set of regional node stresses.

[0057] Based on the real-time data set of regional node stresses generated in the previous step, which contains the stress values of key nodes in each region at consecutive time steps during the simulation process, the system then processes the stress data within each region to calculate the stress gradient change rate. Specifically, for a certain node within a specific region, select its stress values at two consecutive time steps (for example, time points and , where the time interval ), which are and respectively. Calculate the difference between these two stress values, and then divide this difference by the corresponding time interval to obtain the stress gradient change rate of this node within this time period, that is, . This calculation is performed for all representative nodes within the region, and the stress gradient change rates of these nodes may be averaged or the maximum value may be taken to represent the stress gradient change situation of the entire region. Subsequently, the system compares the stress gradient change rate calculated for each region with a preset stress gradient change rate threshold one by one for each region. The setting of this threshold is based on the statistical analysis of the stress change rate of the track structure under normal operating conditions and the understanding of the characteristics of sharp stress changes under abnormal working conditions (such as sudden fracture of track components, abnormal impact of trains, etc.). For example, by analyzing historical monitoring data and simulation cases, it is found that under normal circumstances, the stress change rate of a certain key point on the rail generally does not exceed 500 MPa / s, while in the event of emergencies such as weld cracking or fastener failure, the stress change rate may soar to more than 2000 MPa / s within a short period of time. Therefore, a stress gradient change rate threshold of, for example, 1000 MPa / s can be set as the judgment criterion. When the stress gradient change rate calculated for a certain region (for example, the maximum stress gradient change rate of the nodes within the region reaches 1500 MPa / s) exceeds this preset limit of 1000 MPa / s, the system then identifies this region as an abnormal region and summarizes these identified abnormal regions to form a list of identified abnormal stress regions.

[0058] Based on the list of identified abnormal stress regions generated in the previous step, which indicates which regions have stress change rates beyond the normal range during the current simulation process, potentially signaling sudden changes in structural behavior or the rapid development of potential risks, the system will execute an adaptive model adjustment and re-simulation strategy for each identified abnormal region in this list. Specifically, if an abnormal region was originally configured as an orbital element model with parameter compression configuration (i.e., a low-fidelity model), the system will automatically upgrade it to an orbital element model with full parameter configuration (i.e., a high-fidelity model), and will re-instantiate a model for this region that includes more detailed geometry, more accurate material constitutive relations, and more complex contact conditions. If an abnormal region was originally already configured as an orbital element model with full parameter configuration, but its stress gradient change rate still exceeds the preset limit, this may indicate that even the high-fidelity model fails to fully capture the complex mechanical behavior of this region, or there are local stress concentrations that cannot be accurately resolved. In this case, the system will perform parameter refinement on the full parameter configuration model of this region. Parameter refinement includes further densifying the finite element mesh of this region (e.g., reducing the element size by 50%), using higher-order element types, and introducing more refined material nonlinear models. After completing the model upgrade or parameter refinement for all abnormal regions, based on the adjusted hybrid fidelity model configuration, the latest load conditions and material parameters will be reloaded, and dynamic stress analysis and calculation will be performed again to obtain updated stress data. The finally obtained stress data and related analysis results after adaptive adjustment and recalculation will generate the adaptive simulation output results.

Claims

1. A digital twin simulation system for the maintenance cycle of sleeper tracks, characterized in that, The system includes: A multi-source data integration module that integrates the point cloud data of a laser scanner and the image sequences of UAV photogrammetry, extracts the rail profile point coordinates of the crossties and the geometric parameters of the crosstie positions, and generates an integrated structure attribute set; A geometric consistency calibration module that, based on the preset geometric form of the digital twin model and the integrated structure attribute set, calculates the spatial displacement vector between the reference points of the corresponding rail components in the digital twin model and the measured coordinates, obtains a geometric deviation metric element, and adjusts the rotation matrix of the corresponding rail nodes and crosstie components in the digital twin model according to the geometric deviation metric element to map the geometric form changes and generate a calibrated geometric representation; A topological structure evolution module that retrieves the calibrated geometric representation and, in combination with the maintenance records during the crosstie maintenance period, determines the changes in the component connection relationships caused by crosstie replacement or fastener addition and subtraction, forms a connection change instruction set, and reconstructs the constraints between the rails and crossties according to the connection change instruction set to construct an evolved topological network diagram; An adaptive collaborative simulation module that receives the evolved topological network diagram and the simulation analysis requirements for the crosstie maintenance period, selects a matching track unit model according to the identified stress areas, combines them into a hybrid fidelity model configuration, and performs simulation operations based on the hybrid fidelity model configuration to obtain an adaptive simulation output result.

2. The digital twin simulation system for the maintenance cycle of sleeper tracks according to claim 1, wherein, The steps for obtaining the integrated structure attribute set are as follows: Based on the rail point cloud data collected by the laser scanner and the image sequences obtained by UAV aerial photography, extract the three-dimensional coordinates of the rail edge profile points and the geometric position parameters of the center points of the crosstie cross-sections in each frame. After uniformly projecting them into the WGS-84 coordinate system, align all the acquisition timestamps, and match them one by one through the index corresponding frame sequence to form a set of joint feature point pairs, and obtain an aligned list of joint feature point pairs; Calculate the rail profile curve registration factor according to the aligned list of joint feature point pairs; According to the rail profile curve registration factor and the three-dimensional offset value of each pair of feature points, screen out the point pairs with an offset value greater than the rail profile curve registration factor, recalculate the remapping vector in the global coordinate system and perform coordinate replacement, and at the same time perform coordinate fitting and correction on the geometric parameters of the center points of the crosstie cross-sections. Combine the remapped point coordinates and the corrected geometric parameters to generate an integrated structure attribute set.

3. The digital twin simulation system for the maintenance cycle of pillow tracks according to claim 1, wherein The steps for obtaining the geometric deviation metric element are as follows: Based on the preset standard geometric forms of the rails and crossties in the digital twin model, call the rail edge profile point coordinates and the geometric parameters of the center points of the crosstie cross-sections in the integrated structure attribute set, and match the reference points of the corresponding rail components and the center points of the crosstie components in the digital twin model to form a set of spatial point pairs with one-to-one correspondence between the model and the measured points; Based on the coordinates of each pair of reference points of the rail components and the measured rail edge profile points in the spatial point pairs, calculate the coordinate differences of each point pair along the X, Y, and Z axes in the three-dimensional space one by one, synthesize a complete three-dimensional spatial displacement vector and record the vector length, and at the same time calculate the average spatial displacement vector length of the point pair set to generate a set of rail spatial displacement vectors; Based on the geometric parameters of the center points of each pair of sleeper components and the center points of the measured sleeper cross-sections in the spatial point pairs, extract the Euler angle difference values of the rotational attitude changes between the two points, respectively record the attitude angle differences in the pitch angle, yaw angle, and roll angle, form a set of sleeper attitude angle differences, and combine the set of rail spatial displacement vectors with the set of sleeper attitude angle differences to obtain a geometric deviation metric element.

4. The digital twin simulation system for the maintenance cycle of pillow tracks according to claim 1, characterized in that, The steps for obtaining the calibrated geometric representation are as follows: According to the geometric deviation metric element, extract each spatial displacement vector in the set of rail spatial displacement vectors, respectively analyze the displacement components of the spatial displacement vector on the three coordinate axes of X, Y, and Z. For each corresponding rail node coordinate in the digital twin model, add the numerical values of the displacement components of each axis analyzed to the original coordinate values of the node to form an adjusted rail node coordinate sequence; Based on the adjusted rail node coordinate sequence, call the set of sleeper attitude angle differences in the geometric deviation metric element, analyze the pitch angle difference, yaw angle difference, and roll angle difference of each sleeper in the set one by one, and recalculate the rotation matrix parameters of each sleeper component to generate a corrected sleeper rotation matrix sequence; Based on the adjusted rail node coordinate sequence and the corrected sleeper rotation matrix sequence, apply the rail node coordinate sequence and the sleeper rotation matrix sequence to the original geometric configuration of the rails and sleepers in the digital twin model one by one, and perform a three-dimensional shape mapping update again to form a calibrated geometric representation.

5. The digital twin simulation system for the maintenance cycle of pillow tracks according to claim 1, characterized in that The steps for obtaining the connection change instruction set are as follows: Retrieve the rail node coordinates and sleeper component rotation matrices in the calibrated geometric representation, and item by item identify the adjacency information and connection type identifiers in the original connection relationships between each node and each sleeper. Combine the text content of the maintenance records during the rail-sleeper maintenance cycle, extract all record statements with keywords such as "replacement", "removal", and "addition", and analyze the associated sleeper numbers and fastener numbers to form an initial list of structural connection changes; Based on the initial list of structural connection changes, compare the analyzed sleeper numbers and fastener numbers with the current connection status in the calibrated geometric representation according to the time period and spatial position interval corresponding to each sleeper number. If there is a sleeper number in the list but not in the calibrated geometric representation, it is determined as a newly added sleeper connection. If there is a fastener number in the calibrated geometric representation but not in the list, it is determined as a fastener removal, and a list of component connection relationship changes is generated; According to the list of component connection relationship changes, construct a standardized instruction text format for all connection relationships marked as newly added or removed, and sequentially supplement the instruction type, target number, operation node index, and influence range identifier to generate a connection change instruction set.

6. The digital twin simulation system for the maintenance cycle of pillow tracks according to claim 1, characterized in that The steps for obtaining the evolved topological network diagram are as follows: Based on the connection change instruction set, parse the target number, operation node index, and influence range identifier in each connection change instruction in sequence. Call the rail node coordinates and sleeper component index in the calibration geometric representation body, and determine whether the target number corresponds to the existing node set. If it is a new addition instruction, add this number to the node set and construct an initial connection entry. If it is a removal instruction, delete all connection entries associated with this number to obtain a structure connection reconstruction node table; According to the structure connection reconstruction node table, scan the connection status between each rail node coordinate and the corresponding sleeper component, extract all node pairs with physical connection relationships, and calculate the connection orientation attribute according to the rotation matrix status of the sleeper component to form a constraint relationship reconstruction entry table; Based on the constraint relationship reconstruction entry table, establish an adjacency list structure with the node number as the main index one by one. Embed the connection status, connection direction, and connection label as edge attributes into the corresponding adjacency entries, and synchronously construct an edge index table to generate an evolutionary topology network diagram.

7. The digital twin simulation system for the maintenance cycle of pillow tracks according to claim 1, wherein The steps for obtaining the hybrid fidelity model configuration are as follows: Based on the node structure relationship in the evolutionary topology network diagram, extract the connection edge set between all node pairs, and identify the node numbers and spatial distribution positions associated with each connection edge one by one. Match the structure fatigue monitoring parameter table given in the demand for the simulation analysis of the rail-sleeper maintenance cycle, and screen the regional edge set where the connection edge falls into the fatigue risk identification interval to generate a stress region boundary index list; Based on the stress region boundary index list, calculate the track unit model adaptation factor for each stress region; Based on the track unit model adaptation factor corresponding to each stress region, compare the track unit model adaptation factor with the adaptation value reference threshold item by item. If the track unit model adaptation factor corresponding to the region is greater than or equal to the adaptation value reference threshold, designate the region as a track unit model with complete parameter configuration. Otherwise, designate it as a track unit model with parameter compression configuration, and integrate the two types of track unit models into the hybrid fidelity model configuration of the overall track system.

8. The digital twin simulation system for the maintenance cycle of pillow tracks according to claim 1, characterized in that The steps for obtaining the adaptive simulation output result are as follows: Based on each region delimited by the hybrid fidelity model configuration, call the track unit model with complete parameter configuration or parameter compression configuration for each region one by one, load the current load conditions and material parameters of each track unit model in this region respectively, perform dynamic force analysis and calculation, and obtain the node stress data sequence of each region in real time to generate a regional node stress real-time data set; According to the regional node stress real-time data set, calculate the difference in stress numerical values within consecutive time steps in each region and divide it by the corresponding time interval to obtain the stress gradient change rate of each region. Compare the calculated stress gradient change rate with the preset stress gradient change rate threshold for each region one by one, identify the abnormal regions where the change rate exceeds the preset limit, and form an abnormal stress region identification list; Based on the list of identified abnormal stress regions, for each abnormal region in the list, upgrade the track unit model originally configured with parameter compression configuration to a complete parameter configuration, refine the parameters of the track unit model originally configured with a complete parameter configuration, update the stress data again, and generate an adaptive simulation output result.

Images