A multi-source data fusion broken rail real-time monitoring hierarchical architecture system and method

By constructing a hierarchical architecture system that integrates multi-source data, the problems of single-sensor and insufficient real-time performance in track break monitoring of rail transit systems have been solved, enabling real-time monitoring and prediction of track breaks and improving monitoring accuracy and the level of intelligent operation and maintenance management.

CN123117780APending Publication Date: 2026-11-17CHINA RAILWAY SIYUAN SURVEY & DESIGN GRP CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202611268308.3
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-08-20
Publication Date
2026-11-17

AI Technical Summary

Technical Problem

Existing track break monitoring technologies in rail transit systems suffer from limitations such as single sensing methods, centralized data processing leading to insufficient real-time performance, and a lack of hierarchical collaboration mechanisms, making it difficult to achieve multi-dimensional information acquisition and rapid early warning.

Method used

A hierarchical architecture system for multi-source data fusion is constructed, including a field layer, a station layer, and a central layer. Through multi-source data fusion, edge computing, and digital twin technology, real-time monitoring and prediction of track breakage are achieved.

Benefits of technology

It enables real-time monitoring and prediction of track breakage, improving the accuracy and reliability of monitoring, identifying potential risks in advance, and enhancing the level of intelligent operation and maintenance management and the ability to ensure the safety of track operation.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN123117780A_ABST
    Figure CN123117780A_ABST
Patent Text Reader

Abstract

This invention provides a multi-source data fusion-based hierarchical architecture system and method for real-time monitoring of track breaks. The method includes: a field layer continuously collecting track status data and determining the risk status of track breaks at monitoring nodes; a station layer synchronously receiving track status data uploaded by the field layer in its respective area and calculating the regional risk status in real time, and constructing a regional risk assessment model to comprehensively quantify the risk of track breaks; and a central layer synchronously receiving data uploaded by each station and updating it to a digital twin model, constructing maintenance decision indicators based on risk, trend, and probability, and ranking each section to determine maintenance priorities. This invention constructs a three-level hierarchical architecture of "field layer - station layer - central layer," with each layer forming a closed loop through data transmission and status feedback. Through multi-source data fusion, edge computing, and digital twin technology, it achieves real-time monitoring and prediction of track breaks.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This invention relates to the field of safety monitoring technology for rail transit infrastructure, and more specifically, to a hierarchical architecture system and method for real-time monitoring of track breaks based on multi-source data fusion. Background Technology

[0002] Rail fracture (rail breakage) is one of the most serious structural safety hazards in rail transit systems. Current rail breakage monitoring technologies still face the following key challenges in engineering applications: The limited range of sensing methods makes it difficult to acquire multi-dimensional information. Existing monitoring methods mostly rely on track circuits or single-type sensors, which mainly reflect the electrical or vibration state of the track. They are difficult to simultaneously acquire information on internal track defects, structural strain, and vibration response, resulting in insufficient perception of the early evolution of track breakage and difficulty in timely identification of potential risks.

[0003] Data processing is centralized, but the system lacks real-time performance. Traditional systems generally adopt a centralized architecture of "on-site data acquisition - central processing". Data needs to be uploaded to the center for unified analysis, which has problems such as large transmission delay and insufficient real-time response capability. In the event of a sudden rail break or anomaly, it is difficult to complete the early warning response in a timely manner.

[0004] The lack of a hierarchical collaboration mechanism results in low processing efficiency. In the existing system, there is a lack of clear functional division among field equipment, stations, and the central system, resulting in insufficient multi-level collaboration capabilities, excessively long data processing links, and the inability to quickly identify local anomalies. Therefore, it is necessary to propose a hierarchical architecture system and method for multi-source data fusion, which can achieve efficient perception, hierarchical processing and trend prediction of track derailment through composite sensing, edge computing and digital twin technology. Summary of the Invention

[0005] This invention addresses the technical problems existing in the prior art by providing a hierarchical architecture system and method for real-time monitoring of track breaks based on multi-source data fusion. By constructing a three-level hierarchical architecture of "site layer - station layer - center layer", each layer forms a closed loop through data transmission and status feedback. Through multi-source data fusion, edge computing and digital twin technology, real-time monitoring and prediction of track breaks can be achieved.

[0006] According to a first aspect of the present invention, a hierarchical architecture system for real-time monitoring of track breakage based on multi-source data fusion is provided, comprising: The on-site layer is responsible for real-time data collection and edge analysis of rail breakage risk. The station level is responsible for regional data aggregation, regional risk assessment, and local early warning. The central layer is responsible for overall data management, digital twin mapping, and trend prediction.

[0007] Optionally, the field layer, station layer, and center layer form a hierarchical collaborative processing architecture, with the three layers operating synchronously and collaborating at each level.

[0008] According to a second aspect of the present invention, a method for real-time monitoring of track breakage based on multi-source data fusion is provided, comprising: The on-site layer continuously collects track status data, determines the risk status of track breakage at monitoring nodes, generates on-site risk event E1, and uploads it to the station layer; The station level receives the on-site risk event E1 uploaded by the on-site level of its area, calculates the regional risk status, constructs a regional risk assessment model, comprehensively quantifies the track breakage risk, and generates regional risk event E2. The central layer receives regional risk events E2 uploaded by each station layer and updates them to the digital twin model. Based on the risks, trends, and probabilities, it constructs maintenance decision indicators, sorts each section, and generates a maintenance task list according to the sorting results and distributes it to the operation and maintenance platform.

[0009] Based on the above technical solution, the present invention can also be improved as follows.

[0010] Optionally, the track status data includes: ultrasonic detection data, distributed optical fiber detection data, and track dynamic vibration data; the track status data continuously acquired at the field layer includes: Composite monitoring units are deployed along the track. Each monitoring node integrates an ultrasonic probe array and a distributed fiber optic sensor. The ultrasonic probes are used to detect internal defects in the track, including cracks and fractures, while the distributed fiber optic sensors are used to acquire track strain and vibration response information.

[0011] Optionally, the on-site risk event E1 includes a segment ID, timestamp, comprehensive status index, and risk level; the regional risk event E2 includes a region ID, risk level, risk index, and trend information.

[0012] Optionally, the determination of the risk status of track breakage at the monitoring node includes: Multi-source features are normalized and fused to construct a comprehensive state index. ; Based on comprehensive status indicators The orbital state is classified and judged: if the comprehensive state index Less than the abnormal threshold If the overall status indicators are normal, then the status is normal. Greater than the abnormal threshold Less than the rail break risk threshold If the overall state index is abnormal, then the facial expression is abnormal; if the overall state index is abnormal, then the facial expression is abnormal. Greater than the rail breakage risk threshold This indicates a risk of track breakage.

[0013] Optional, anomaly threshold and rail breakage risk threshold It is determined by a combination of historical sample statistics, on-site calibration, and dynamic correction.

[0014] Optionally, the formula for constructing the comprehensive state index is expressed as:

[0015] In the formula, It is represented as a comprehensive state index, dimensionless, and used to characterize the overall health status of the orbit. This is expressed as the average amplitude of the ultrasonic echo; This is represented as the propagation time offset; Expressed as track strain, it is used to reflect the stress state of the track structure; Represented as the root mean square value of vibration; This is represented as the baseline propagation time; Represented as a reference amplitude; Represented as reference strain; This is represented as a reference vibration value; These are weighting coefficients, which are dimensionless.

[0016] Optionally, the construction of a regional risk assessment model to comprehensively quantify the risk of rail failure includes: The status indicators of each node are weighted and fused to obtain the overall status of the region; A regional risk assessment model was constructed to comprehensively quantify the risk of rail breakage and obtain the final risk index. Based on the magnitude of risk indicators, the status of the region is classified. If the final risk indicator is less than the abnormal threshold, it indicates normal registration; if the final risk indicator is greater than the abnormal threshold but less than the high-risk threshold, it indicates an early warning level; if the final risk indicator is greater than the high-risk threshold, it indicates a high-risk level.

[0017] Optionally, the step of constructing maintenance decision indicators based on risks, trends, and probabilities, ranking each section, and thus determining maintenance priorities and distributing them to the operation and maintenance platform includes: The regional status data uploaded by each station are aggregated to form a global input dataset; Construct a digital twin model to uniformly represent the operational status of the line; A track health index is introduced to achieve global health assessment, a regional risk evolution model is established to predict the development trend of track status, and a track break probability model is introduced to quantify risk into decision-making indicators. Maintenance decision indicators are constructed based on risks, trends, and probabilities. Each section is then ranked. The higher the decision indicator, the higher the probability of a rail breakage accident, the faster the risk increases, or the greater the probability of a rail breakage. This has a more significant impact on the safety of line operation and therefore a higher maintenance priority.

[0018] The technical effects and advantages of this invention are as follows: This invention provides a hierarchical architecture system and method for real-time monitoring of track breaks based on multi-source data fusion. By employing multi-source data fusion, edge computing, and digital twin technologies, it achieves real-time monitoring and prediction of track breaks. This invention constructs a three-level hierarchical architecture of "site layer—station layer—central layer," with each layer forming a closed loop through data transmission and status feedback, thus realizing real-time monitoring and prediction of track breaks. Attached Figure Description

[0019] Figure 1 The flowchart illustrates a hierarchical architecture method for real-time monitoring of track breakage using multi-source data fusion, as provided in this embodiment of the invention. Detailed Implementation

[0020] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0021] Understandably, given the deficiencies in the background technology, this invention proposes a hierarchical architecture system for real-time monitoring of track breakage based on multi-source data fusion, specifically as follows: Figure 1 As shown, it includes: The on-site layer is responsible for real-time data collection and edge analysis of rail breakage risk. The station level is responsible for regional data aggregation, regional risk assessment, and local early warning. The central layer is responsible for overall data management, digital twin mapping, and trend prediction.

[0022] During operation, the three layers do not work independently, but rather operate synchronously and collaborate step by step. The three belong to a hierarchical collaborative processing architecture, rather than a simple serial processing architecture. For example, for a 50km long track, the entire section is controlled by the central layer (usually in the line's integrated maintenance center), which oversees 5 stations (i.e., the station layer), and each station manages 10km (corresponding to the field layer).

[0023] The on-site layer deploys composite monitoring units along the track, with each monitoring node integrating an ultrasonic probe array and a distributed fiber optic sensor. The ultrasonic probes are used to detect internal defects in the track (such as cracks and fractures), while the distributed fiber optic sensors are used to acquire track strain and vibration response information, thereby achieving multi-dimensional perception of the track's structural condition.

[0024] A station level typically has 5 to 20 monitoring nodes (depending on the line length and monitoring density), with each node corresponding to an independent status indicator. .

[0025] As the decision-making core of the entire system, the central layer summarizes and analyzes the regional status data uploaded by each station level along the entire line, and realizes global mapping and trend prediction of the operating status of the entire track system through a digital twin model, thereby supporting risk warning and operation and maintenance decision-making at the entire line level.

[0026] On the other hand, embodiments of the present invention provide a real-time monitoring method for rail breakage using multi-source data fusion in the above-mentioned system, comprising the following steps: The on-site layer continuously collects track status data and determines the risk status of track breakage at monitoring nodes, generating an on-site risk event E1 and uploading it to the station layer; the on-site risk event E1 includes section ID, timestamp, comprehensive status index, and risk level; The station level receives the on-site risk event E1 uploaded by the on-site level of its area, calculates the regional risk status, constructs a regional risk assessment model, comprehensively quantifies the track break risk, and generates regional risk event E2; the regional risk event E2 includes regional ID, risk level, risk index and trend information. The central layer receives regional risk events E2 uploaded by each station layer and updates them to the digital twin model. Based on the risks, trends, and probabilities, it constructs maintenance decision indicators, sorts each section, and generates a maintenance task list according to the sorting results and distributes it to the operation and maintenance platform.

[0027] By adopting the above technical solutions, this embodiment of the invention constructs a three-level hierarchical architecture of "site layer - station layer - center layer". Each layer forms a closed loop through data transmission and status feedback. Through multi-source data fusion, edge computing and digital twin technology, real-time monitoring and prediction of track derailment are achieved.

[0028] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the following description is provided in conjunction with the accompanying drawings. Figure 1 Further details are provided through examples.

[0029] Site layer Composite monitoring units are deployed along the track, with each monitoring node integrating an ultrasonic probe array and a distributed fiber optic sensor. The ultrasonic probes are used to detect internal defects in the track (such as cracks and fractures), while the distributed fiber optic sensors are used to acquire track strain and vibration response information, thereby achieving multi-dimensional perception of the track's structural condition.

[0030] The on-site layer first synchronously collects track status data, which includes: ultrasonic detection data, distributed fiber optic detection data, and track dynamic vibration data; the collected data can be expressed by the following formula:

[0031] In the formula, X represents the set of multi-source monitoring feature vectors collected at the field layer, which is used to characterize the comprehensive operating status of the current track node; Expressed as ultrasonic echo amplitude, unit: V, used to reflect the reflection intensity of internal defects in the track; Expressed as ultrasonic propagation time, in μs, used to characterize changes in the internal structure of the orbit; Expressed as track strain, unit: με, used to reflect the stress state of the track structure; It is expressed as vibration velocity, with the unit: mm / s, and is used to reflect the dynamic response of the track.

[0032] In the inspection of broken rails or welds, the amplitude of the ultrasonic signal reflects the reflection intensity of the internal structure. This is achieved by calculating the average amplitude. This allows us to obtain the overall echo level of the node or channel, which can be used to make a preliminary judgment on internal defects or anomalies in the track, such as cracks, gaps, or inclusions.

[0033] After acquiring the raw data, statistical processing is performed on the ultrasound signals to extract the mean amplitude characteristics of the raw ultrasound signals:

[0034] In the formula, Expressed as the average amplitude of the ultrasonic echo, in units of V; This is expressed as the number of sampling points; It is expressed as the ultrasonic echo amplitude at the kth sampling point (unit: V).

[0035] Subsequently, to reflect changes in the sound propagation path, a propagation time offset is introduced:

[0036] In the formula, The value is expressed as propagation time offset, in μs; Tu represents the current measured ultrasound propagation time, in μs. The reference propagation time is expressed in μs. Used to characterize the effect of changes in the internal structure of the track on the propagation path of sound waves.

[0037] Furthermore, the vibration signal is processed using root mean square (RMS) analysis to obtain the vibration intensity index:

[0038] In the formula, It is expressed as the root mean square value of vibration, in mm / s; this index is used to characterize the energy level of track vibration. The value represents the vibration velocity at the k-th sampling time, in mm / s; N represents the total number of vibration signal sampling points. After extracting a single feature, the multi-source features are normalized and fused to construct a comprehensive state index:

[0039] in, It is represented as a comprehensive state index, dimensionless, and used to characterize the overall health status of the orbit. Represented as reference amplitude (V); Expressed as track strain, unit: με, used to reflect the stress state of the track structure; Represented as reference strain (με); Expressed as reference vibration value (mm / s); These are weighting coefficients, which are dimensionless.

[0040] The assessment of the risk status of track breakage at monitoring nodes includes: The multi-source features are normalized and fused to construct the above-mentioned comprehensive state index. ; Based on comprehensive status indicators The orbital status is classified and judged in stages:

[0041] in, This is represented as the abnormal threshold; Represented as the rail breakage risk threshold; abnormal threshold and rail breakage risk threshold It is determined by a combination of historical sample statistics, on-site calibration, and dynamic correction.

[0042] When the orbit is in a healthy state, the comprehensive status index SSS is collected within a certain time window, and its mean μ is calculated. S and standard deviation σ S :

[0043] In the formula, μ Srepresents the mean value of the comprehensive state indicator under healthy conditions; M represents the number of historical healthy samples; S m represents the comprehensive state indicator corresponding to the m-th historical sample.

[0044]

[0045] σ S represents the standard deviation of the comprehensive state indicator under healthy conditions; On this basis, the anomaly threshold and the rail break risk threshold are respectively determined as:

[0046]

[0047] where a and b are threshold adjustment coefficients, and 0<a<b. In practical engineering, a can be 2 to 3, and b can be 4 to 6. The anomaly threshold is used to identify a significant deviation of the rail state from the healthy reference, and the rail break risk threshold is used to identify a severe abnormal state with rail break risk. The anomaly threshold S y and the rail break risk threshold S d are determined by a combination of historical sample statistics, on-site calibration and dynamic correction.

[0048] Further, the results of manual inspection, historical rail break or crack sample data can be combined to and perform correction. When the on-site false alarm rate is high, appropriately increase a or b; when the risk of missed alarm is high, appropriately decrease a or b. Thus, dynamic updating of the thresholds with changes in line conditions, seasonal temperature, train operation density and equipment state is realized.

[0049] Through the above threshold determination method, the present invention can avoid the problem of insufficient adaptability caused by the use of fixed empirical thresholds, so that the anomaly threshold and the rail break risk threshold can be automatically calibrated according to the actual operation state of different line sections, thereby improving the accuracy and reliability of rail break risk identification.

[0050] For example, when a certain node detects an increase in ultrasonic amplitude, a propagation time delay and enhanced vibration, the calculated exceeds the threshold , the system determines that it is in a rail break risk state and triggers an alarm.

[0051] Station layer The station layer centrally processes data from multiple on-site monitoring nodes in its affiliated section, so as to realize comprehensive assessment and risk determination of the rail state in the regional scope.

[0052] In actual engineering projects, 5 to 20 monitoring nodes are typically set up in a section under the jurisdiction of a station (the number is determined according to the length of the line and the monitoring density), and each node corresponds to an independent status indicator. .

[0053] First, the status indicators of each node are weighted and fused to obtain the overall regional status:

[0054] in, It is expressed as a regional comprehensive status index and is dimensionless. This is represented as the status index of the i-th node; These are expressed as weighting coefficients; the determination of weights mainly includes the importance of the node's location (e.g., higher weight in turnout areas); historical failure frequency (higher weight in high-risk areas); and sensor reliability (higher weight in stable nodes).

[0055] After obtaining the regional average state, a spatial dispersion index is introduced to avoid misjudging individual outliers:

[0056] in, It is expressed as spatial dispersion, dimensionless; used to reflect the degree of difference in the state of each node. The smaller the value, the more consistent the anomaly (more likely to be a real risk); the larger the value, the more likely it is a local misjudgment or noise.

[0057] Furthermore, to analyze the dynamic change trend of the regional state, the rate of state change is calculated:

[0058] in, Expressed as rate of change, unit: 1 / s; The time interval is expressed in seconds (s); t represents the current sampling time. Based on this, a regional risk assessment model is constructed to comprehensively quantify the risk of railroad malfunction, resulting in a final risk index, expressed as:

[0059] Where R represents the final risk indicator; S s This is represented as the overall regional status; Expressed as the rate of change; This is expressed as spatial dispersion; These are weighting coefficients, which are dimensionless.

[0060] The status of the region is classified according to the magnitude of the risk indicators:

[0061] When the regional risk or rate of change reaches a threshold, the reporting mechanism is triggered: .

[0062] In the formula, R y Represented as the abnormal threshold; R g This is indicated as a high-risk threshold; λ represents the rate of change; λ represents the threshold of the rate of change.

[0063] Example description For example, within a certain station's jurisdiction, there are 5 monitoring nodes in a certain section. The status values ​​of 3 of these nodes have significantly increased. The overall status of the area is calculated. rate of change Final risk indicators Exceeding the high-risk threshold The station system determined that there was a concentrated anomaly in the section and sent an early warning message to the central layer.

[0064] central layer As the decision-making core of the entire system, the central layer summarizes and analyzes the regional status data uploaded by each station level along the entire line, and realizes global mapping and trend prediction of the operating status of the entire track system through a digital twin model, thereby supporting risk warning and operation and maintenance decision-making at the entire line level.

[0065] First, the regional status data uploaded by each station are aggregated to form a global input dataset:

[0066] in, Let M represent the comprehensive state index (dimensionless) of the j-th region at time t; M represents the number of regions at the station level.

[0067] Based on this, a digital twin model is constructed to uniformly represent the operational status of the line:

[0068] in, This is expressed as the rate of state change of the j-th region (unit: 1 / s); It is expressed as a regional risk index (dimensionless).

[0069] To achieve a comprehensive health assessment, a line health index is introduced:

[0070] in, Represented as the global health index (value range 0-1), when When the value is close to 1, it indicates that the track condition of the entire line is good. A decrease indicates an increase in the number of risky zones or a worsening of the risk level.

[0071] Furthermore, to predict the development trend of orbital states, a regional risk evolution model is established:

[0072] in, Predict the risk value for region j at the next time step; Expressed as changing acceleration, with units of 1 / s²; changing acceleration The formula is expressed as: , When A j >0 indicates that regional risk is increasing rapidly, A j A value less than 0 indicates that the region is stable and at risk. Represented as weight parameters; To quantify risk into a decision-making indicator, a track break probability model (nonlinear mapping function) is introduced:

[0073] in, This represents the probability of a track break (0 to 1). R0 represents the sensitivity coefficient; R0 represents the risk benchmark value. ; Finally, maintenance decision indicators are constructed based on risks, trends, and probabilities to rank each section and thus determine the maintenance priority. The larger the value, the higher the maintenance priority for that area.

[0074]

[0075] in, The maintenance decision index (dimensionless) for the j-th monitoring section is used to characterize the overall maintenance priority of that section. This represents the regional risk index (dimensionless) of the j-th monitoring segment, reflecting the risk level of the current orbital state. This represents the rate of risk change in the j-th monitoring segment, expressed in 1 / s, and is used to characterize the trend of risk growth or decline. This represents the probability of a rail break occurring in the j-th monitoring section, with a value ranging from 0 to 1. This represents the risk weighting coefficient (dimensionless), used to measure the importance of the current risk level in maintenance decisions; The trend weight coefficient (dimensionless) is used to measure the degree of impact of risk change trends on maintenance decisions. The value represents the probability weighting coefficient (dimensionless), used to measure the impact of the probability of a rail break on maintenance decisions; j represents the monitoring section number in the line, j=1,2, M; M represents the total number of monitoring sections in the line.

[0076]

[0077] in: =0.4 0.6, =0.2 0.3, =0.2 0.3.

[0078] This is to highlight the leading role of the current risk status in maintenance decisions.

[0079] Maintain decision indicators The larger the value, the higher the probability of a rail breakage accident in that section, the faster the risk increases, or the greater the probability of a rail breakage, and the more significant the impact on the safety of line operation. Therefore, its maintenance priority is higher.

[0080] The system sorts all sections in descending order based on the maintenance decision indicators corresponding to them:

[0081] The system generates a maintenance task list based on the sorting results. The top 10% of sections are designated as Level 1 maintenance targets, the top 10% to 30% of sections are designated as key inspection targets, and the remaining sections are designated as routine inspection targets.

[0082] In summary, this invention proposes a hierarchical architecture system and method for real-time monitoring of track breaks based on multi-source data fusion. By constructing a three-tiered architecture of "site layer—station layer—central layer," and utilizing multi-source data fusion, edge computing, and digital twin technology, real-time monitoring and prediction of track breaks are achieved. A closed loop is formed between each layer through data transmission and status feedback.

[0083] Compared with the prior art, the present invention has the following advantages and implementation effects: The combination of multi-source fusion and a hierarchical architecture significantly improves the accuracy and reliability of track break monitoring. This invention constructs a multi-source composite sensing system consisting of an ultrasonic probe array and distributed fiber optic sensors, and combines this with a hierarchical data fusion mechanism at the field, station, and central levels to achieve a step-by-step processing flow from single-point sensing to comprehensive regional judgment and then to global analysis. Compared to traditional single-sensor or single-layer analysis methods, this effectively reduces the risk of misjudgment caused by environmental interference and local anomalies, significantly improving the accuracy of track break monitoring and the overall reliability of the system.

[0084] This invention introduces risk trend prediction and probabilistic modeling to achieve proactive early warning of track conditions. Based on a regional risk index and its changing trends, a risk prediction model is constructed. Through probabilistic mapping, the risk is quantified into the probability of track breakage, transforming track condition assessment from "static detection" to "dynamic prediction." Compared to existing technologies that rely solely on current detection values, this approach can identify potential risks in advance, enabling preventative maintenance and improving the safety of track operation.

[0085] This invention constructs a unified quantitative decision-making system and digital twin model to improve the intelligence level of operation and maintenance management. By establishing a unified quantitative system of risk index, track break probability, and operation and maintenance priority indicators, and combining it with a central-layer digital twin model, it achieves dynamic mapping and visual analysis of the entire line's status. Compared to operation and maintenance methods relying on manual experience, it can automatically prioritize risks across multiple sections and optimize resource allocation, improving the scientific nature, efficiency, and intelligence of operation and maintenance decisions.

[0086] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.

[0087] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0088] Although preferred embodiments of the invention have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including both the preferred embodiments and all changes and modifications falling within the scope of the invention.

[0089] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of the claims of this invention and their equivalents, this invention also intends to include these modifications and variations.

[0090] Finally, it should be noted that the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A hierarchical architecture system for real-time monitoring of track breakage using multi-source data fusion, characterized in that, include: The on-site layer is responsible for real-time data collection and edge analysis of rail breakage risk. The station level is responsible for regional data aggregation, regional risk assessment, and local early warning. The central layer is responsible for overall data management, digital twin mapping, and trend prediction.

2. The hierarchical architecture system for real-time monitoring of track breakage based on multi-source data fusion according to claim 1, characterized in that, The field layer, station layer, and center layer form a hierarchical collaborative processing architecture, with the three layers operating synchronously and collaborating at each level.

3. A method for real-time monitoring of track breaks based on multi-source data fusion, used in the hierarchical architecture system for real-time monitoring of track breaks based on multi-source data fusion as described in claim 1, characterized in that, Includes the following steps: The on-site layer continuously collects track status data, determines the risk status of track breakage at monitoring nodes, generates on-site risk event E1, and uploads it to the station layer; The station level receives the on-site risk event E1 uploaded by the on-site level of its area, calculates the regional risk status, constructs a regional risk assessment model, comprehensively quantifies the track breakage risk, and generates regional risk event E2. The central layer receives regional risk events E2 uploaded by each station layer and updates them to the digital twin model. Based on the risks, trends, and probabilities, it constructs maintenance decision indicators, sorts each section, and generates a maintenance task list according to the sorting results and distributes it to the operation and maintenance platform.

4. The method for real-time monitoring of track breakage based on multi-source data fusion according to claim 3, characterized in that, The track status data includes: ultrasonic detection data, distributed optical fiber detection data, and track dynamic vibration data; the track status data continuously collected by the field layer includes: Composite monitoring units are deployed along the track. Each monitoring node integrates an ultrasonic probe array and a distributed fiber optic sensor. The ultrasonic probe is used to detect internal defects in the track, including cracks and fractures. The distributed fiber optic sensor is used to acquire track strain and vibration response information.

5. The method for real-time monitoring of track breakage based on multi-source data fusion according to claim 3, characterized in that, The determination of the risk status of track breakage at the monitoring node includes: Multi-source features are normalized and fused to construct a comprehensive state index. ; Based on comprehensive status indicators The orbital state is classified and judged: if the comprehensive state index Less than the abnormal threshold If the overall status indicators are normal, then the status is normal. Greater than the abnormal threshold Less than the rail break risk threshold If the overall status index is negative, it indicates an abnormal state; if the overall status index is negative, it indicates an abnormal state. Greater than the rail breakage risk threshold This indicates a risk of track breakage.

6. The method for real-time monitoring of track breakage based on multi-source data fusion according to claim 5, characterized in that, Abnormal threshold and rail breakage risk threshold It is determined by a combination of historical sample statistics, on-site calibration, and dynamic correction.

7. The method for real-time monitoring of track breakage based on multi-source data fusion according to claim 5, characterized in that, The construction of comprehensive status indicators The formula is expressed as: In the formula, It is represented as a comprehensive state index, dimensionless, and used to characterize the overall health status of the orbit. This is expressed as the average amplitude of the ultrasonic echo; This is represented as the propagation time offset; Expressed as track strain, it is used to reflect the stress state of the track structure; Represented as the root mean square value of vibration; This is represented as the baseline propagation time; Represented as a reference amplitude; Represented as reference strain; Represented as a reference vibration value; These are weighting coefficients, which are dimensionless.

8. The method for real-time monitoring of track breakage based on multi-source data fusion according to claim 3, characterized in that, The construction of the regional risk assessment model, which comprehensively quantifies the risk of rail failure, includes: The status indicators of each node are weighted and fused to obtain the overall status of the region; A regional risk assessment model was constructed to comprehensively quantify the risk of rail breakage based on the overall regional condition, and the final risk index was obtained. The status of a region is classified according to the magnitude of the final risk indicator. If the final risk indicator is less than the abnormal threshold, it indicates normal registration; if the final risk indicator is greater than the abnormal threshold but less than the high-risk threshold, it indicates an early warning level; if the final risk indicator is greater than the high-risk threshold, it indicates a high-risk level.

9. The method for real-time monitoring of track breakage based on multi-source data fusion according to claim 3, characterized in that, The process of constructing maintenance decision indicators based on risks, trends, and probabilities, ranking each section, and generating a maintenance task list based on the ranking results and distributing it to the operation and maintenance platform includes: The regional status data uploaded by each station are aggregated to form a global input dataset; Construct a digital twin model to uniformly represent the operational status of the line; A track health index is introduced to achieve global health assessment, a regional risk evolution model is established to predict the development trend of track status, and a track break probability model is introduced to quantify risk into decision-making indicators. Maintenance decision indicators are constructed based on risks, trends, and probabilities. Each section is then ranked. The higher the decision indicator, the higher the probability of a rail breakage accident, the faster the risk increases, or the greater the probability of a rail breakage. This has a more significant impact on the safety of line operation and therefore a higher maintenance priority.

10. The method for real-time monitoring of track breakage based on multi-source data fusion according to claim 3, characterized in that, The on-site risk event E1 includes the segment ID, timestamp, comprehensive status index, and risk level; the regional risk event E2 includes the region ID, risk level, risk index, and trend information.