Fault determination method and device based on network rail tunnel system and train

Abstract

The invention provides a fault determination method based on a network rail tunnel system. The fault determination method can be applied to the technical field of rail vehicle operation and management. The fault determination method comprises the following steps: acquiring first real-time running state information of a target train in a target time period and first real-time running environment information of a running route; determining a target association relationship among the real-time operation state information, the real-time pantograph-catenary monitoring data, the real-time track monitoring data and the real-time tunnel monitoring data based on a predetermined map; and inputting the target association relationship into a trained target map neural network model, and outputting potential fault information related to the operation environment of the target train, the potential fault information at least comprising a potential fault point and a fault occurrence probability, so that before the target train travels to the potential fault point, troubleshooting operation is performed on the potential fault point.

Description

Technical Field

[0001] The present disclosure relates to the technical field of operation and management of rail vehicles, and more particularly, to a fault determination method, device and train based on a catenary-track-tunnel system. Background Art

[0002] In the traditional electromechanical system integration mode, trains, track maintenance, etc. are independent of each other, lacking overall planning and management, resulting in data islands among trains, catenary, tracks, and tunnels, and insufficient fault collaborative processing capabilities. At the same time, when a fault occurs, the cause and scope of the fault are determined based on the respective data of trains, track maintenance, etc., relying on manual or preset algorithms to diagnose the occurred fault phenomena. It is difficult to give early warnings for potential faults and impossible to formulate targeted prevention and emergency plans in advance. Summary of the Invention

[0003] In view of this, the present disclosure provides a fault determination method, device, train, electronic device, storage medium and computer program product based on a catenary-track-tunnel system.

[0004] One aspect of the present disclosure provides a fault determination method based on a catenary-track-tunnel system, including: obtaining first real-time operation state information of a target train in a target time period and first real-time operation environment information of the line on which the train travels, where the first real-time operation environment information includes real-time catenary monitoring data, real-time track monitoring data, and real-time tunnel monitoring data collected by train monitoring devices and corresponding to the same space-time as the real-time operation state information; determining a target association relationship among the real-time operation state information, the real-time catenary monitoring data, the real-time track monitoring data, and the real-time tunnel monitoring data based on a predetermined map, where the predetermined map is constructed according to historical catenary data, historical track data, and historical tunnel data and is used to represent the association relationship among the historical catenary data, the historical track data, and the historical tunnel data; inputting the target association relationship into a trained target graph neural network model, and outputting potential fault information related to the operation environment of the target train, where the potential fault information at least includes potential fault points and the probability of fault occurrence, so as to perform fault elimination operations on the potential fault points before the target train travels to the potential fault points.

[0005] According to an embodiment of the present disclosure, a predetermined graph is constructed by the following method: obtaining historical operation state information and historical operation environment information of a target train within a historical time period, where the historical operation state information includes historical driving stability data, and the real-time operation environment information includes historical pantograph-catenary data, historical track data, and historical tunnel data collected by train monitoring devices at the same time and space as the historical operation state information; constructing at least one transaction dataset according to L data types included in the historical driving stability data, historical pantograph-catenary data, historical track data, and historical tunnel data, where at least one transaction dataset includes L transactions corresponding to the L data types; determining a plurality of frequent item sets according to the support degree of each transaction and the support degree threshold; generating a plurality of frequent sub-item sets corresponding to the frequent item sets according to each transaction in each frequent item set; calculating the lift of each frequent sub-item set, and determining the association relationship of each transaction in the frequent sub-item set according to the lift; constructing a predetermined graph of each transaction according to the association relationship of each transaction.

[0006] According to an embodiment of the present disclosure, the L data types include: the historical driving stability data includes at least one of the following data types or a combination of multiple data types: smooth detection data of the target train, instability detection data, vibration detection data, wheel spin and skid data; the historical pantograph-catenary data includes at least one of the following data types or a combination of multiple data types: catenary voltage data of the target train, catenary current data, pantograph-catenary contact pressure data, arcing rate data of the line, pantograph-catenary foreign object data; the historical track data includes at least one of the following data types or a combination of multiple data types: rail profile data, rail corrugation data, turnout abnormal state data, gauge detection data, rail surface state data during wheel spin and skid; the historical tunnel data includes at least one of the following data types or a combination of multiple data types: tunnel foreign object intrusion data, tunnel clearance data, tunnel video anomaly data.

[0007] According to an embodiment of the present disclosure, the target graph neural network model is trained by the following method: obtaining a predetermined graph and historical fault diagnosis information within a historical time period; using the predetermined graph as historical sample data and the historical fault diagnosis information as labels to optimize the initial graph neural network model to obtain the target graph neural network model.

[0008] According to an embodiment of the present disclosure, the fault determination method based on the pantograph-track-tunnel system further includes: determining a target fault point from potential fault points according to the second real-time operation state information, the second real-time operation environment information, and the fault occurrence probability of multiple reference trains in the same area, so as to perform fault troubleshooting on the target fault point in advance.

[0009] According to an embodiment of the present disclosure, the potential fault information further includes the potential fault type and the expected occurrence time of the potential fault. The method for determining a fault based on the network-rail-tunnel system further includes: determining the severity and the impact scope of the potential fault according to the potential fault point, the potential fault type, and the expected occurrence time of the potential fault; and determining the processing priority and the processing strategy of the potential fault according to the severity and the impact scope.

[0010] Another aspect of the present disclosure provides a fault determination device based on a network-rail-tunnel system, including: a first acquisition module, configured to acquire first real-time operation state information of a target train in a target time period and first real-time operation environment information of the line on which the train travels, where the first real-time operation environment information includes real-time pantograph-catenary monitoring data, real-time track monitoring data, and real-time tunnel monitoring data collected by train monitoring devices and corresponding to the same space-time as the real-time operation state information; a first determination module, configured to determine a target association relationship among the real-time operation state information, the real-time pantograph-catenary monitoring data, the real-time track monitoring data, and the real-time tunnel monitoring data based on a predetermined graph, where the predetermined graph is constructed according to historical pantograph-catenary data, historical track data, and historical tunnel data and is used to represent the association relationship among the historical pantograph-catenary data, the historical track data, and the historical tunnel data; and an output module, configured to input the target association relationship into a trained target graph neural network model and output potential fault information related to the operation environment of the target train, where the potential fault information includes at least a potential fault point and a fault occurrence probability, so as to perform a fault elimination operation on the potential fault point before the target train travels to the potential fault point.

[0011] Another aspect of the present disclosure provides a train, including: a vehicle body; a braking member; and a fault determination device based on a network-rail-tunnel system.

[0012] Another aspect of the present disclosure provides an electronic device, including: one or more processors; and a memory, configured to store one or more programs, where when the one or more programs are executed by the one or more processors, the one or more processors are caused to implement the method as described above.

[0013] Another aspect of the present disclosure provides a computer-readable storage medium storing computer-executable instructions, where the instructions, when executed, are used to implement the method as described above.

[0014] Another aspect of the present disclosure provides a computer program product including computer-executable instructions, where the instructions, when executed, are used to implement the method as described above.

[0015] According to an embodiment of the present disclosure, by collecting real-time operation status information and operation environment information of a target train, determining the correlation relationship between various data in combination with a predetermined graph spectrum, and analyzing using a trained graph neural network model, potential fault information can be output in advance, including potential fault points and the probability of fault occurrence. Thus, the staff can take measures before the train reaches the potential fault point, mitigate the impact of the fault, and even avoid the occurrence of the fault. Among them, the predetermined graph spectrum is constructed based on historical data, which can make full use of all historical data to mine the potential correlation between different data, determine the target correlation relationship of real-time data based on this, provide a more accurate relationship for fault prediction, and improve the accuracy of the fault determination method. Further, on the one hand, timely obtaining potential fault information and performing fault troubleshooting operations in advance can effectively reduce the risk of faults occurring during the train operation, ensure the safe operation of the train, reduce train delays, outages, etc. caused by faults, improve the reliability and stability of railway transportation, and ensure the safety of passengers' lives and property. On the other hand, according to the potential fault information, the staff can reasonably arrange resources such as maintenance personnel, equipment, and spare parts, avoiding waste and over-allocation of resources. Making targeted preparations before the fault occurs can ensure that resources can be put into use in a timely manner when needed, improving the utilization efficiency of resources. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Through the following description of the embodiments of the present disclosure with reference to the drawings, the above and other objects, features, and advantages of the present disclosure will become clearer. In the drawings:

[0017] Figure 1 Schematically shows an exemplary system architecture to which the fault determination method based on the network-rail-tunnel system according to an embodiment of the present disclosure can be applied;

[0018] Figure 2 Schematically shows a flowchart of the fault determination method based on the network-rail-tunnel system according to an embodiment of the present disclosure;

[0019] Figure 3 Shows a schematic diagram of the method for determining faults based on multiple trains according to an embodiment of the present disclosure;

[0020] Figure 4 Schematically shows a block diagram of a fault determination device based on the network-rail-tunnel system according to an embodiment of the present disclosure;

[0021] Figure 5 Schematically shows a block diagram suitable for implementing the fault determination method based on the network-rail-tunnel system according to an embodiment of the present disclosure. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0022] Hereinafter, embodiments of the present disclosure will be described with reference to the accompanying drawings. However, it should be understood that these descriptions are merely exemplary and are not intended to limit the scope of the present disclosure. In the following detailed description, for the sake of explanation, numerous specific details are set forth to provide a comprehensive understanding of the embodiments of the present disclosure. However, obviously, one or more embodiments can also be implemented without these specific details. In addition, in the following description, descriptions of well-known structures and technologies are omitted to avoid unnecessarily confusing the concepts of the present disclosure.

[0023] The terms used herein are merely for describing specific embodiments and are not intended to limit the present disclosure. The terms "including", "comprising", etc. used herein indicate the presence of the described features, steps, operations, and / or components, but do not exclude the presence or addition of one or more other features, steps, operations, or components.

[0024] All terms used herein (including technical and scientific terms) have the meanings commonly understood by those skilled in the art, unless otherwise defined. It should be noted that the terms used herein should be interpreted as having a meaning consistent with the context of this specification and should not be interpreted in an idealized or overly rigid manner.

[0025] In cases where expressions similar to "at least one of A, B, and C, etc." are used, generally, it should be interpreted according to the meaning commonly understood by those skilled in the art (for example, "a system having at least one of A, B, and C" should include, but not be limited to, a system having only A, only B, only C, having A and B, having A and C, having B and C, and / or having A, B, and C, etc.).

[0026] In the embodiments of the present disclosure, in terms of the collection, update, analysis, processing, use, transmission, provision, disclosure, storage, etc. of the involved data (for example, including but not limited to user personal information), they all comply with the provisions of relevant laws and regulations, are used for legal purposes, and do not violate public order and good customs. In particular, necessary measures are taken for user personal information to prevent illegal access to user personal information data and to safeguard the security of user personal information, network security, and national security.

[0027] In the embodiments of the present disclosure, the authorization or consent of the user is obtained before obtaining or collecting user personal information.

[0028] In the traditional electromechanical system integration mode, the operation data of trains, the status monitoring of tracks, and the environmental parameters of tunnels are scattered, lacking an effective data interaction and fusion mechanism. When a fault occurs, it is difficult to quickly locate the root cause of the fault from a global perspective, and the collaborative disposal efficiency is low.

[0029] Current fault diagnosis relies on manual experience or preset algorithms, and can only conduct post - mortem analysis on the occurred fault phenomena. This passive fault handling method is difficult to capture the subtle abnormal changes during the system operation. It cannot identify the early signs of potential faults, let alone predict the possibility and scope of influence of fault occurrence. In the complex and changeable railway operation environment, potential faults may evolve into major safety accidents at any time, while the existing mechanism cannot formulate targeted prevention and emergency plans in advance. Once a fault occurs, it is extremely easy to trigger chain reactions such as train delays and equipment damage, which not only brings huge economic losses to railway transportation enterprises, but also seriously threatens the lives of passengers and the stable operation of railway transportation.

[0030] Embodiments of the present disclosure provide a fault determination method based on a catenary - track - tunnel system, including: obtaining first real - time operation state information of a target train in a target time period and first real - time operation environment information of the traveled line, where the first real - time operation environment information includes real - time catenary monitoring data, real - time track monitoring data, and real - time tunnel monitoring data collected by train monitoring equipment in the same space - time as the real - time operation state information; determining a target correlation relationship among the real - time operation state information, the real - time catenary monitoring data, the real - time track monitoring data, and the real - time tunnel monitoring data based on a predetermined map, where the predetermined map is constructed according to historical catenary data, historical track data, and historical tunnel data and is used to represent the correlation relationship among the historical catenary data, the historical track data, and the historical tunnel data; inputting the target correlation relationship into a trained target graph neural network model and outputting potential fault information related to the operation environment of the target train, where the potential fault information at least includes potential fault points and the probability of fault occurrence, so as to perform fault exclusion operations on the potential fault points before the target train travels to the potential fault points.

[0031] Figure 1 Exemplary system architecture 100 to which the fault determination method based on the catenary - track - tunnel system can be applied according to an embodiment of the present disclosure is schematically shown. It should be noted that Figure 1 What is shown is only an example of the system architecture to which the embodiments of the present disclosure can be applied, to help those skilled in the art understand the technical content of the present disclosure, but it does not mean that the embodiments of the present disclosure cannot be used in other devices, systems, environments or scenarios.

[0032] As Figure 1 shown, the system architecture 100 according to this embodiment includes a catenary - track - tunnel system 110 and a train monitoring system 120.

[0033] The catenary - track - tunnel system 110 includes a catenary detection system 111, a track detection system 112, a tunnel monitoring system 113, and a host 114.

[0034] The pantograph-catenary detection system 111 is used to obtain catenary parameter data, pantograph status data, pantograph-catenary relationship data, etc. collected by the pantograph-catenary detection equipment installed on the top of the train. For example, detection of conductor height and pull-out value, pantograph-catenary arcing detection, temperature detection, hard point detection, etc.

[0035] The track detection system 112 is used to obtain track geometric parameter data, track component status data, train operation environment data, etc. collected by the track detection equipment installed at the bottom of the train. Among them, the track detection system 112 also includes: a track status inspection device, a track geometric parameter detection device, a rail profile detection device, etc. (not shown in the figure). The track status inspection device can obtain the line image information collected by the track detection equipment at high speed, and extract damage information such as scratches, fish-scale injuries, spalling, wave wear on the rail surface, crack and spalling of the ballast bed, missing fasteners, etc. from the image; use a high-precision camera to dynamically collect the track surface, and then use digital image processing technology to automatically identify and transmit real-time faults for fasteners, rail surface, and ballast bed defects. The track geometric parameter detection device can establish a navigation coordinate system to realize relative positioning and attitude measurement of the carrier. At the same time, combined with a vibration compensation device, vibration compensation and gauge measurement are realized; combined with a positioning synchronization system, precise positioning synchronization is realized. Finally, track geometric parameter information such as gauge, superelevation, alignment, and vertical profile is output in real time. The rail profile detection device can perform real-time, dynamic, and high-precision detection on the entire cross-section and wear of the rail. Automatically establish a standard rail profile, and match and analyze each collected profile with the standard profile, and output the vertical and horizontal wear results in real time.

[0036] The tunnel monitoring device 113 is used to obtain tunnel clearance data, tunnel structure status data, etc. collected by the tunnel detection equipment installed at the front of the train or at the gangway position. Specifically, a high-speed laser scanner can be used to quickly scan the tunnel surface to obtain three-dimensional information of the tunnel cross-section, presenting the spatial form of the tunnel. Compare the tunnel cross-section data collected at high speed with a preset subway clearance model (specifying standards such as the dimensions that the subway tunnel space must meet) to judge whether the actual tunnel space meets the standard, obtain the tunnel clearance detection result, and check for problems such as intrusion. Identify diseases on the tunnel surface, such as cracks and spalling, from the collected two-dimensional images to master the tunnel health status.

[0037] The main unit 114 is usually installed in the car cabinet or under the seat, and is used to collect, collate, analyze, determine faults and give early warnings, and store data collected by the pantograph-catenary detection device 111, the track detection device 112, and the tunnel monitoring device 113.

[0038] The train monitoring device 130 monitors the train operation status in real time, such as the train speed, running direction, etc.; the train equipment status, such as the traction system, braking system, bogie, doors, etc.; it also includes: train number information, train running kilometers, train fault diagnosis information, train signal equipment PHM (Prognostics and Health Management) data and related diagnosis information. Among them, the train PHM diagnosis data can be used to guide the daily train inspection behavior.

[0039] The control center 120 includes a track maintenance management system 121, a train dispatching system 122, other relevant subsystems, and terminal equipment (not shown in the figure). The track maintenance management system 121 is used to receive the track, tunnel and catenary data transmitted by the track, tunnel and catenary system 110, and classify, store and analyze these data for subsequent query and use. The train dispatching system 122 is used to monitor the train position and status in real time, formulate and adjust the train operation strategy, and coordinate train meeting and avoidance, etc. The terminal equipment provides a visual operation interface to display the data processing and analysis results, warning prompt information, etc. to the track maintenance personnel, so that the track maintenance personnel can grasp the vehicle and environment information in real time to make a plan for potential faults.

[0040] The train monitoring device 130 monitors the train operation status in real time, such as the train speed, running direction, etc.; the train equipment status, such as the traction system, braking system, bogie, doors, etc.; it also includes: train number information, train running kilometers, train fault diagnosis information, train signal equipment PHM (Prognostics and Health Management) data and related diagnosis information. Among them, the train PHM diagnosis data can be used to guide the daily train inspection behavior.

[0041] The positioning and synchronization system 140 obtains the train position information by means of radio frequency tags installed at specific positions on the track or satellite positioning.

[0042] Figure 2 The flowchart of the fault determination method based on the track, tunnel and catenary system according to an embodiment of the present disclosure is schematically shown.

[0043] As Figure 2 shown, the method includes operations S210~S230.

[0044] In operation S210, obtain the first real-time operation status information of the target train in the target time period and the first real-time operation environment information of the line traveled, where the first real-time operation environment information includes the real-time pantograph-catenary monitoring data, real-time track monitoring data and real-time tunnel monitoring data collected by the train monitoring equipment in the same space-time as the real-time operation status information.

[0045] According to an embodiment of the present disclosure, the target train may be the train currently in operation. The first real-time operation status information is used to reflect the real-time operation condition of the target train itself, such as whether the train runs smoothly, whether it is unstable, whether it vibrates, the speed and acceleration of the train, etc. The first operation environment information is used to reflect the environmental state of the line on which the target train operates, and at least includes real-time pantograph-catenary monitoring data, real-time track monitoring data, and real-time tunnel monitoring data. Among them, the real-time pantograph-catenary monitoring data reflects the working state of the pantograph-catenary system, such as the voltage, current, contact pressure, etc. of the catenary; the real-time track monitoring data, such as the geometric dimensions, wear conditions, and vibrations of the track; the real-time tunnel monitoring data, such as the temperature, humidity, and structural deformation in the tunnel.

[0046] In operation S220, based on a predetermined map, determine the target correlation relationship among the real-time operation status information, the real-time pantograph-catenary monitoring data, the real-time track monitoring data, and the real-time tunnel monitoring data, where the predetermined map is constructed according to historical pantograph-catenary data, historical track data, and historical tunnel data, and is used to characterize the correlation relationship among the historical pantograph-catenary data, historical track data, and historical tunnel data.

[0047] According to an embodiment of the present disclosure, the predetermined map is constructed according to historical pantograph-catenary data, historical track data, and historical tunnel data, and may also include the historical full-scale operation status information of the historical full-scale trains collected by the train monitoring system. Among them, the historical pantograph-catenary data, historical track data, historical tunnel data, and historical full-scale operation status information serve as the nodes of the map. The correlation relationships among the various nodes are connected by edges. For example, there is an edge of "power transmission" relationship between the vehicle and the pantograph-catenary, an edge of "physical contact" relationship between the vehicle and the track, an edge of "spatial position" relationship between the vehicle and the tunnel, an edge of "indirect influence" relationship between the pantograph-catenary and the track, and so on. Each edge also has attribute information, which is used to describe the strength, frequency, confidence, etc. of the relationship. For example, for the "physical contact" relationship edge of "vehicle-track", its attributes may include the degree of influence of the track condition on the wear of vehicle components. Through the predetermined map, the mutual connections among the real-time data can be mined, so as to more comprehensively understand the interaction among various factors during the train operation process.

[0048] In operation S230, input the target correlation relationship into the trained target graph neural network model, and output the potential fault information related to the operation environment of the target train, where the potential fault information at least includes potential fault points and the probability of fault occurrence, so as to perform fault troubleshooting operations on the potential fault points before the target train travels to the potential fault points.

[0049] According to an embodiment of the present disclosure, the target graph neural network is a neural network for processing data with a graph structure, which can learn and analyze the relationships between data. The trained model can output potential fault information related to the operating environment of the target train according to the input target association relationship. The potential fault information at least includes potential fault points, that is, the positions where faults may occur, such as section A of the track, and the corresponding fault occurrence probability, such as the probability of skidding when the train travels to section A of the track is 76%. Thus, by obtaining potential fault information in advance, it is possible to perform targeted fault removal operations on potential fault points before the target train reaches the potential fault points, thereby preventing the occurrence of faults and ensuring the safe operation of the train.

[0050] According to an embodiment of the present disclosure, by collecting the real-time operating status information and operating environment information of the target train, determining the association relationship between each data in combination with a predetermined graph spectrum, and using the trained graph neural network model for analysis, it is possible to output potential fault information in advance, including potential fault points and fault occurrence probabilities. Thus, the staff can take measures before the train reaches the potential fault points, reduce the impact brought by the faults, and even avoid the occurrence of faults. Among them, the predetermined graph spectrum is constructed based on historical data, which can make full use of the full amount of historical data to mine the potential associations between different data, determine the target association relationship of real-time data based on this, provide a more accurate relationship for fault prediction, and improve the accuracy of the fault determination method. Further, on the one hand, timely obtaining potential fault information and performing fault removal operations in advance can effectively reduce the risk of faults occurring during the train operation, ensure the safe operation of the train, reduce the train delays, outages, etc. caused by faults, improve the reliability and stability of railway transportation, and ensure the life and property safety of passengers. On the other hand, according to the potential fault information, the staff can reasonably arrange resources such as maintenance personnel, equipment, and spare parts, avoiding waste and over-allocation of resources. Making targeted preparations before the occurrence of faults can ensure that resources can be put into use in a timely manner when needed, improving the utilization efficiency of resources.

[0051] According to an embodiment of the present disclosure, in addition to the graph neural network model mentioned in the present disclosure, a support vector machine (SVM), a random forest, and other neural network models can also be used, and training can be performed using sample data and labels. The vehicle operation state data and the track, network, and tunnel data collected in real time are input into the trained model, and the real-time data is predicted and analyzed according to the learned correlation relationship to determine whether each data point belongs to the normal range. For example, for the track elevation data, if the model predicts that the data at a certain measurement point significantly deviates from the normal elevation range, it is marked as a potential abnormal data point. Cluster analysis is performed on the identified abnormal data points, and adjacent abnormal data points are clustered into potential problem areas according to the spatial position (such as the mileage position of the track) and time correlation of the data points. For example, if a continuous number of measurement points in a certain section of the track are marked as abnormal and these abnormal data points are also continuous in time (such as occurring within the same time period), then this track section is marked as a potential problem area. At the same time, according to the type and severity of the abnormal data, a priority is assigned to each problem area for subsequent maintenance and processing. The marked potential problem areas and their related information (such as location, abnormal type, priority, etc.) are output.

[0052] According to an embodiment of the present disclosure, the method for constructing a predetermined graph includes steps 11 to 16.

[0053] Step 11: Obtain the historical operation state information and historical operation environment information of the target train during a historical time period. Among them, the historical operation state information includes historical driving stability data, and the real-time operation environment information includes historical pantograph-catenary data, historical track data, and historical tunnel data collected by the train monitoring equipment at the same time and space as the historical operation state information.

[0054] According to an embodiment of the present disclosure, the historical time period can be any time period before constructing the preset graph. When setting the historical time period, in order to make the collected data more comprehensive, all factors that affect the train and the environment should be covered as much as possible (such as weather conditions, holiday information, passenger flow, etc.). Generally, it is considered that the longer the duration of the historical time period is set, the more comprehensive the data collection is.

[0055] According to an embodiment of the present disclosure, vehicle operation data is obtained from a train monitoring system, including but not limited to vehicle real-time operation speed, acceleration, ride quality index, instability state, vibration data, etc.); data such as contact force, current, voltage, and pantograph slider wear of the pantograph-catenary is collected from pantograph-catenary monitoring equipment; geometric parameters of the track (such as gauge, vertical irregularity, alignment deviation), rail profile, and corrugation data are collected through track detection equipment; data such as temperature, humidity, ventilation state, and lining structure state of the tunnel are obtained using tunnel detection equipment. At the same time, the timestamps and spatial location information of each data collection are recorded to ensure the spatio-temporal consistency of the data.

[0056] According to an embodiment of the present disclosure, relevant data when faults occur in vehicles, pantograph-catenary, tracks, and tunnels are collected from historical data, including fault types, fault occurrence times, fault locations, and operation data and environmental data for a period of time before and after the faults occur. The fault data and normal operation data are marked and distinguished to form a complete data set.

[0057] According to an embodiment of the present disclosure, noise data, outliers, and missing values in the data are removed. Data of different types and different dimensions are unified and normalized to eliminate the influence of dimension differences between data on subsequent analysis.

[0058] Step 12, at least one transaction data set is constructed according to L types of data included in historical driving stability data, historical pantograph-catenary data, historical track data, and historical tunnel data, where at least one transaction data set contains L transactions corresponding to L types of data.

[0059] According to an embodiment of the present disclosure, historical driving stability data may include M types of data, at least including: ride detection data, instability detection data, vibration detection data, and wheel spin / slide data of the target train.

[0060] According to an embodiment of the present disclosure, historical pantograph-catenary data may include N types of data, at least including: catenary voltage data, catenary current data, pantograph-catenary contact pressure data, arcing rate data of the line, and pantograph-catenary foreign object data of the target train.

[0061] According to an embodiment of the present disclosure, historical track data may include X types of data, at least including: rail profile data, rail corrugation data, turnout abnormal state data, gauge detection data, and rail surface state data during wheel spin / slide.

[0062] According to an embodiment of the present disclosure, historical tunnel data may include Y types of data, at least including: tunnel foreign object intrusion data, tunnel clearance data, and tunnel video anomaly data.

[0063] Among them, M, N, X, and Y are all positive integers greater than or equal to 1, and M + N + X + Y = L, that is, a transaction dataset includes at least 4 transactions corresponding to 4 data types.

[0064] According to an embodiment of the present disclosure, at least one transaction dataset further includes status categories corresponding to L transactions. The status category corresponding to a transaction is used to characterize the degree represented by the data value of this type of data. Exemplarily, for vehicle speed data, the status category can be divided into several intervals such as low speed, medium speed, and high speed; for track irregularity data, the status category is divided into categories such as slight, moderate, and severe according to its severity.

[0065] For ease of understanding, the status categories of the above data are exemplarily divided into three levels.

[0066] According to an embodiment of the present disclosure, the smoothness detection data (excellent / good / poor), where excellent indicates that all smoothness indicators during the train operation are within the ideal range, without obvious shaking, bumping, etc.; good indicates that there are slight instability signs, but they do not exceed the specified threshold; poor indicates that the smoothness indicators are significantly abnormal, and the train shakes and bumps violently.

[0067] According to an embodiment of the present disclosure, the instability detection data (low / medium / high), where low indicates that the detected instability signal is weak, and the possibility of the train experiencing an instability condition is extremely low. Medium indicates that there is a certain degree of instability signal, and the train may experience slight instability. High indicates a strong instability signal.

[0068] According to an embodiment of the present disclosure, the vibration detection data (weak / medium / strong), where weak indicates that the train vibration amplitude is small, within the normal vibration range, and will not cause adverse effects on train equipment and passengers. Medium indicates that the vibration amplitude has increased, exceeding the normal range but not reaching the severe level, and may have a certain impact on some precision equipment. Strong indicates that the train vibrates violently, which may cause equipment loosening, damage, and even affect the operation safety and stability of the train.

[0069] According to an embodiment of the present disclosure, the wheel spin and skid data (weak / medium / strong), where weak indicates that the wheel spin and skid degree of the train is slight, and has little impact on the train running speed and power output. Medium indicates that the wheel spin and skid are relatively obvious and have already had a certain impact on the train operation. Strong indicates that the wheel spin and skid are severe, the train power drops significantly, the running speed decreases significantly, and may even cause the train to stop.

[0070] According to an embodiment of the present disclosure, the catenary voltage data (low / medium / high), where low indicates that the catenary voltage is lower than the lower limit of the normal working range, which may cause insufficient train power. Medium indicates that the catenary voltage is within the normal fluctuation range. High indicates that the catenary voltage is higher than the upper limit of the normal working range, which may cause damage to train electrical equipment.

[0071] According to an embodiment of the present disclosure, network current data (low / medium / high), where "low" indicates a relatively small current intensity, representing that the train is in a light load state or there is an abnormality in the electrical system, resulting in insufficient power output. "Medium" indicates that the network current is at a normal level, meeting the power consumption requirements under the current operating conditions of the train. "High" indicates an excessive current, which may be caused by reasons such as an overloaded train, a malfunction of electrical equipment, or an abnormality in the catenary.

[0072] According to an embodiment of the present disclosure, pantograph-catenary contact pressure data (low / medium / high), where "low" indicates an excessively small contact pressure, which may lead to poor contact between the pantograph and the catenary, resulting in phenomena such as off-line and arcing. "Medium" indicates a moderate contact pressure, which can ensure good contact between the pantograph and the catenary. "High" indicates an excessively large contact pressure, which will increase the wear of the pantograph and the catenary, shorten the service life of the equipment, and may even damage the pantograph and the catenary.

[0073] According to an embodiment of the present disclosure, arcing rate data of the line (low / medium / high), where "low" indicates that arcing occurs less frequently, the pantograph-catenary system is in good operating condition, and the loss of equipment is small. "Medium" indicates that arcing occurs at a certain frequency, and its development trend needs to be concerned, and maintenance and adjustment should be carried out in a timely manner. "High" indicates that arcing occurs frequently, which will accelerate the wear of the pantograph and the catenary, reduce the reliability of the equipment, and may cause power supply failures.

[0074] According to an embodiment of the present disclosure, foreign object data of the pantograph-catenary (low / medium / high), where "low" indicates that the detected foreign objects have little impact on the pantograph-catenary system, or there are almost no foreign objects. "Medium" indicates that there are a certain number or volume of foreign objects, which may interfere with the normal contact of the pantograph-catenary. "High" indicates that there are many foreign objects or the volume is large, seriously threatening the safety of the pantograph-catenary system, and may cause damage to the pantograph and catenary failures.

[0075] According to an embodiment of the present disclosure, rail profile data (excellent / good / poor), where "excellent" means very close to the standard profile and the rail wear is small; "good" means there is a certain degree of wear or deformation, but it does not affect the normal operation of the train; "poor" means that the profile change is large, which may affect the running stability and safety of the train.

[0076] According to an embodiment of the present disclosure, rail corrugation data (severe / medium / light), where "light" means the degree of corrugation is light and has little impact on the train operation; "medium" means the degree of corrugation is moderate and may cause a certain degree of vibration of the train; "severe" means the corrugation is severe, which will significantly affect the running smoothness of the train and may even lead to increased damage to the wheels and rails.

[0077] According to the embodiments of the present disclosure, the abnormal status data of the turnout (none / light / heavy) are provided, wherein "none" indicates that the turnout is in normal status and can complete functions such as switching and locking normally; "light" indicates that there are some minor problems with the turnout, such as slight wear of components, indication errors, etc., but do not affect the basic functions of the turnout and the passage of trains; "heavy" indicates that there are serious problems with the turnout, such as the point rail not fitting tightly, the switch machine failure, etc., which may cause the turnout to be unable to be used normally and affect the operation of the train.

[0078] According to the embodiments of the present disclosure, the track gauge detection data (standard / near standard / deviate from standard), wherein "standard" means that the track gauge fully complies with the prescribed standard value and can provide a good running track for the train; "near standard" means that the track gauge has a certain deviation from the standard value, but within the permitted range, the impact on the train operation is small; "deviate from standard" means that the track gauge deviation exceeds the permitted range, which may cause the train to run unsteadily or even have the risk of derailment.

[0079] According to the embodiments of the present disclosure, the track surface status data (dry / wet / oily) during idling and sliding, among which, the "dry" track surface can provide better adhesion conditions, which is helpful for the train to resume normal operation; the "wet" track surface will reduce the adhesion coefficient and increase the possibility of idling and sliding; the "oily" track surface will seriously reduce the adhesion coefficient, aggravate the idling and sliding of the train, and it is not easy to resume normal operation.

[0080] According to the embodiments of the present disclosure, tunnel foreign object intrusion limit data (none / a small amount / a large amount) is provided, wherein "none" indicates that no foreign object has intruded the limit in the tunnel; "a small amount" indicates that there are some small foreign objects intruding the limit, but they have not posed a serious threat to the safety of train operation; "a large amount" indicates that there are more or larger foreign objects intruding the limit, which may collide with the train, seriously threatening the safety of train operation.

[0081] According to the embodiments of the present disclosure, tunnel limit data meets / is close to critical / exceeds), "meets" means that the tunnel limit fully meets the requirements for safe train operation; "close to critical" means that the tunnel limit is close to the minimum allowable value, which may have an impact on some oversized freight trains or train operations under special circumstances; "exceeds" means that the tunnel limit exceeds the prescribed range, which may cause dangerous situations such as scratches between the train and the tunnel wall.

[0082] According to the embodiments of the present disclosure, tunnel video abnormality data (none / minor abnormality / serious abnormality), "none" means that the tunnel video surveillance picture is normal and no abnormality is found; "minor abnormality" means that some small abnormalities that do not affect the safety of train operation appear in the video, such as flickering lights, partial blurred images, etc.; "serious abnormality" means that obvious abnormalities appear in the video, such as fire, collapse, etc., which will pose a serious threat to the safety of train operation.

[0083] According to an embodiment of the present disclosure, one or more time windows are obtained from a historical time period. The preprocessed data is divided into multiple transactions according to the time windows. For example, if the historical time period is 2 years, one year, one month, one day, or one hour can be used as a time window. Each transaction contains all data types of physical objects of vehicles, pantograph-catenary systems, tracks, and tunnels within the same time window and the corresponding status categories of these physical objects.

[0084] Exemplarily, when the time window is 12:00 - 17:00, multiple transaction data sets are constructed according to the obtained historical full-scale data.

[0085] Transaction data set 1: [Smoothness (good), network current data (low), pantograph-catenary contact pressure data (low), rail profile data (good)] of vehicle A traveling on line A.

[0086] Transaction data set 2: [Smoothness (good), network current data (low), pantograph-catenary contact pressure data (low), rail profile data (good), turnout anomaly (slight anomaly), and tunnel video anomaly data (none)] of vehicle B traveling on line A.

[0087] Transaction data set 3: [Smoothness (good), network current data (low), pantograph-catenary contact pressure data (low), rail profile data (good), turnout anomaly (slight anomaly), tunnel video anomaly data (none), rail surface state data (dry) during wheel spin and skid] of vehicle C traveling on line A.

[0088] Step 13: Determine multiple frequent item sets according to the support degree of each transaction and the support degree threshold.

[0089] According to an embodiment of the present disclosure, calculate the support degree of each transaction in each transaction data set and compare it with the support degree threshold. If the support degree threshold is set as: in the above transaction data, the number of occurrences reaching 50% of the total number of transaction data sets (3 transaction data sets) (i.e., the number of occurrences is greater than or equal to 2 times) is used as the judgment criterion for frequent item sets (if it is in the form of a grade, assume that the frequent grade threshold for each transaction is "at least reaching the grade of 2 or more occurrences"). The number of occurrences of the transactions "Smoothness (good)", "Network current data (low)", and "Rail profile data (good)" has all reached 3 times, which is greater than or equal to 50% of the total number of transaction data sets (i.e., 2 times), so the item sets containing these transactions are all frequent item sets.

[0090] Step 14: Generate multiple frequent sub-item sets corresponding to the frequent item sets according to each transaction in each frequent item set.

[0091] According to an embodiment of the present disclosure, for a frequent item set, all its non-empty subsets are frequent sub-item sets.

[0092] Simplify the transaction dataset 1: [Stationarity (good), network flow data (low), pantograph-catenary contact pressure data (low), rail profile data (good)].

[0093] Simplify it to: [A (good), B (low), C (low), D (good)].

[0094] Then the frequent subsets of the transaction dataset 1 include: [A (good)], [B (low)], [C (low)], [D (good)], [A (good), B (low)], [B (low), C (low)], [A (good), B (low), C (low), D (good)], etc.

[0095] Step 15, calculate the lift of each frequent subset, and determine the association relationship between each transaction in the frequent subset according to the lift.

[0096] According to the embodiments of the present disclosure, exemplarily, in addition to the transaction dataset 1, there are also transaction dataset 4: [A (excellent), B (high), C (low), D (poor)] and transaction dataset 5: [A (good), B (low), C (high), D (medium)]. Calculate the lift of the frequent subset B (low) and the frequent subset C (low) respectively: In the 3 transaction datasets, B (low) appears 2 times, so P(B (low)) = 2 / 3; In the 3 transaction datasets, C (low) appears 2 times, so P(C (low)) = 2 / 3; The situation where B (low) and C (low) appear simultaneously is in transaction datasets 1 and 5, a total of 2 times, so . The lift of B (low) to C (low) refers to Equation (1):

[0097] Equation (1)

[0098] Where, I represents a frequent subset, J represents another frequent subset, represents the lift of the I frequent subset to the J frequent subset.

[0099] Calculated according to Equation (1) , when the lift is greater than 1, it indicates that and There is a positive correlation between them, that is, when B is in the "low" state, C is more likely to be in the "low" state. In this example, it shows that there is a positive correlation between the network flow data (low) and the pantograph-catenary contact pressure data (low), that is, when the network flow is in the "low" state, the pantograph-catenary contact pressure is more likely to be in the "low" state. This association relationship may reflect that there is a certain problem in the power supply system, resulting in both of them being in a lower state at the same time, which can be used as a reference for detecting abnormalities or faults.

[0100] According to an embodiment of the present disclosure, for each frequent itemset, all possible association rules are generated. By calculating and analyzing the lift of different frequent item sets, the association relationships between various transactions in the full amount of historical network-rail-tunnel data are calculated. The generated association rules are further screened to remove redundant rules and rules that do not have practical significance.

[0101] Step 16 constructs a predetermined graph for each transaction according to the association relationship of each transaction.

[0102] According to an embodiment of the present disclosure, vehicles, pantographs, tracks, and tunnels are used as nodes of the knowledge graph. Each node contains attribute information. For example, the vehicle node contains attributes such as vehicle ID, vehicle type, service life, real-time operating status (speed, acceleration, smoothness, etc.), historical fault records, etc.; the pantograph node contains attributes such as pantograph ID, model, material, real-time contact force, current, voltage, historical maintenance records, etc.; the track node contains attributes such as track ID, laying location, laying material, service life, real-time geometric parameters (gauge, vertical irregularity, etc.), historical maintenance records, etc.; the tunnel node contains attributes such as tunnel ID, structure type, completion time, real-time environmental parameters (temperature, humidity, ventilation, etc.), past accident records, etc. The preprocessed data and the association relationships mined through the above algorithms are added as node attributes to the corresponding nodes. For example, the association rule related to the vehicle (such as "if the vehicle speed is high and the track gauge has a serious deviation, then the vehicle vibration intensifies") is used as an attribute of the vehicle node, recording the specific content and confidence level of the association rule.

[0103] According to an embodiment of the present disclosure, the edges between nodes are determined according to the association rules mined by the above algorithms. The edges represent the association relationships between nodes. For example, there is an "electric power transmission" relationship edge between the vehicle and the pantograph, a "physical contact" relationship edge between the vehicle and the track, a "spatial position" relationship edge between the vehicle and the tunnel, an "indirect influence" relationship edge (influenced by vehicle operation) between the pantograph and the track, an "environmental influence" relationship edge between the pantograph and the tunnel, a "structural interaction" relationship edge between the track and the tunnel, etc. Attributes are added to each edge to describe information such as the strength, frequency, and confidence level of the relationship. For example, for the "physical contact" relationship edge of "vehicle - track", its attributes can include the degree of influence of the track condition on the wear of vehicle components (obtained from historical data statistics), the confidence level of this relationship in the algorithm association rules, etc.

[0104] According to an embodiment of the present disclosure, a professional graph database is selected to store the knowledge graph. At the same time, a real-time update mechanism for the knowledge graph is established. When new operation data or fault data is generated, the attributes of the corresponding nodes and the relationships of the edges are updated in a timely manner. At the same time, the above-mentioned association relationship mining algorithm is periodically re-run to update the association rules, and the knowledge graph is optimized and expanded.

[0105] According to an embodiment of the present disclosure, by calculating the support degree, the frequency of occurrence of each data item of the vehicle, network, track, and tunnel in the historical data can be determined. On this basis, calculating the lift degree can discover the association strength between different data items, and then accurately locate which combinations of vehicle, network, track, and tunnel data are closely related to abnormal situations. For example, when the network flow data is within a certain specific range and at the same time the rail profile data presents a certain feature, it may show a strong association with the abnormal data in the tunnel video, which is beneficial to quickly lock the key factor combinations that may cause abnormalities.

[0106] According to an embodiment of the present disclosure, constructing a knowledge graph based on the mined association relationships can be used to establish a prediction model. When a specific state of a data item is monitored, according to the association relationships, it is possible to predict in advance the possible abnormalities or faults that may occur in the vehicle, network, track, and tunnel systems, so as to issue a warning in a timely manner. For example, if it is found that the network voltage data fluctuates and matches the combination of network flow data and vehicle ride comfort data that have caused abnormal wear of the rail surface in the historical data, measures can be taken in advance, such as strengthening the inspection and maintenance of the rail surface, to avoid the occurrence of faults or reduce the impact of faults.

[0107] According to an embodiment of the present disclosure, the training steps of the target graph neural network model include step 21 and step 22.

[0108] Step 21: Obtain a predetermined graph and historical fault diagnosis information within a historical time period.

[0109] Step 22: Use the predetermined graph as historical sample data and the historical fault diagnosis information as labels to optimize the initial graph neural network model to obtain the target graph neural network model.

[0110] According to an embodiment of the present disclosure, obtain a predetermined graph, and obtain historical fault diagnosis information that records various fault situations occurring in the vehicle, network, track, and tunnel systems within a historical time period, including the type of fault (such as minor fault, medium fault, severe fault), the location where the fault occurs (potential fault point), the probability of the fault occurring, etc. Use the obtained association relationship graph as historical sample data and the historical fault diagnosis information as labels, and input them into the initial graph neural network model. The model makes predictions based on the input association relationship graph, and then compares the prediction results with the labels (historical fault diagnosis information) to calculate the prediction error. Through the backpropagation algorithm, the error is transmitted from the output layer to the input layer to adjust the parameters of the model (such as the weights of nodes, the weights of edges, etc.), so that the prediction results of the model gradually approach the true labels. After multiple iterative trainings, continuously optimize the parameters of the model until the prediction results meet the training termination conditions to obtain the target graph neural network model.

[0111] According to an embodiment of the present disclosure, in a complex environment of a track-tunnel system with multi-factor interactions, traditional fault diagnosis methods often rely on manual experience or simple rule judgments, which not only have low efficiency but also are prone to misjudgments. By using a predetermined graph spectrum as historical sample data and combining historical fault diagnosis information as labels to train an initial graph neural network model, the model can effectively learn the internal relationships between different data associations and faults. When the model faces new operation data, it can more accurately predict potential faults, take preventive measures in advance, and ensure the safe and stable operation of the system. At the same time, the target graph neural network model trained and optimized with a large amount of historical data can quickly analyze the input association relationship graph spectrum and obtain the fault diagnosis result, greatly shortening the fault diagnosis time, improving the diagnosis efficiency, helping to quickly locate the cause of the fault, and reducing the adverse impact on operation. In addition, the predetermined graph spectrum comprehensively describes the complex relationships of various data in the system, enabling the target graph neural network model to better adapt to the characteristics of complex systems, process multi-source heterogeneous data, deeply mine the potential associations between data, accurately capture abnormal conditions during system operation, and significantly improve the fault diagnosis and prediction capabilities for complex systems.

[0112] Figure 3 The schematic diagram of a method for determining faults according to multiple trains according to an embodiment of the present disclosure is shown.

[0113] According to an embodiment of the present disclosure, the fault determination method based on the track-tunnel system further includes: determining a target fault point from potential fault points according to the second real-time operation state information, the second real-time operation environment information, and the fault occurrence probability of multiple reference trains in the same area, so as to perform fault elimination on the target fault point in advance.

[0114] According to an embodiment of the present disclosure, as Figure 3 shown, the track-tunnel data of Train 1, Train 2, and Train 3 are obtained simultaneously and sent to the track maintenance management system. Considering the data of multiple reference trains comprehensively, the fault information at the points is verified and supplemented according to the operation conditions of different trains. For example, if multiple trains all show similar abnormal operation states (such as increased vibration) when passing through a certain area, and the track-tunnel environment information in this area also shows some abnormal indicators (such as an increase in the rail profile deviation), and the fault occurrence probability of the corresponding potential fault point in this area is relatively high, then this potential fault point is more likely to become the target fault point. The train dispatching system generates an operation strategy based on the potential fault information and the track-tunnel data of multiple trains and realizes the dispatching of each vehicle.

[0115] According to an embodiment of the present disclosure, the smoothness detection data of the target vehicle, combined with the detection data of the turnout, track irregularity, etc. of the reference train, can identify abnormal line states such as turnouts in the same section, and perform fusion diagnosis, which can improve the accuracy of locating the abnormal section. At the same time, combined with the train cluster analysis, identify whether it is due to the track or the vehicle, and achieve precise maintenance.

[0116] According to an embodiment of the present disclosure, through the diagnostic result data such as instability alarms or the appearance of instability harmonic characteristics of multiple trains and multiple cars in the same section, combined with the data of the comprehensive inspection vehicle such as rail profile, rail wear, and gauge detection, and using multi-source data for fusion diagnosis, for the phenomenon of lateral instability of the frame, the accuracy of identifying abnormal vehicle states and track states can be improved.

[0117] According to an embodiment of the present disclosure, data such as vehicle network voltage and network current are combined with the detection data such as the static current-carrying capacity of the inspection vehicle, pantograph-catenary contact pressure, and arcing rate of the line for fusion diagnosis. Through various detection means of the inspection vehicle, supplemented by the real-time data of network voltage and network current during the operation of the vehicle, comprehensively judge the current collection performance of the pantograph-catenary and study the pantograph-catenary matching relationship.

[0118] According to an embodiment of the present disclosure, when the current vehicle detects that there is a foreign object intrusion in the tunnel, which affects the train operation safety, immediately feedback the tunnel intrusion data (position, video images before and after the intrusion position) to the control center for confirmation, and take timely fault handling measures to provide guarantee for the subsequent train operation safety.

[0119] According to an embodiment of the present disclosure, based on the real-time sharing of multi-vehicle network-rail-tunnel detection data, perform a fusion situation analysis on the tunnel clearance data detected by multiple vehicles at the same position, analyze whether there is a foreign object gradually approaching the vehicle operation clearance. If so, feedback the tunnel intrusion data (position, video images before and after the intrusion position) to the control center for confirmation, and take timely fault handling measures to provide guarantee for the subsequent train operation safety.

[0120] According to an embodiment of the present disclosure, when the current vehicle detects that there is a foreign object on the pantograph-catenary, which affects the train operation, immediately feedback the pantograph-catenary foreign object data (position, video images before and after the foreign object position) to the control center for confirmation, and take timely fault handling measures to provide guarantee for the subsequent train operation safety. And by optimizing the train control strategy, lower the pantograph in the foreign object area, and then raise the pantograph after sliding over the foreign object position to ensure the operation efficiency under abnormal conditions.

[0121] According to an embodiment of the present disclosure, when the vehicle idles or coasts, the analysis of track detection data is triggered to analyze whether it is caused by wet rail surface, and the track detection data at this position is uploaded to the control center for subsequent fault verification. If it is determined that it is caused by wet rail surface, the data is promptly shared with other vehicles. When other vehicles run to this location, the idling and coasting faults are reduced by reducing the traction force or braking force. At the same time, the relationship between the vehicle braking slip data and the track surface state is determined, such as the influence of water or oil on the track surface on the vehicle braking distance.

[0122] According to an embodiment of the present disclosure, once the target fault point is determined, fault elimination operations can be carried out in advance. It is possible to take corresponding measures to eliminate potential fault hazards before the actual occurrence of the fault, avoid the impact of the fault on the train operation, and ensure the normal operation of the network-rail-tunnel system and the safety of the train.

[0123] According to an embodiment of the present disclosure, the potential fault information further includes the potential fault type and the estimated occurrence time of the potential fault, and the method further includes Step 31 to Step 32.

[0124] Step 31, according to the potential fault point, potential fault type and the estimated occurrence time of the potential fault, determine the severity and the influence range of the potential fault;

[0125] Step 32, according to the severity and the influence range, determine the processing priority and processing strategy of the potential fault.

[0126] According to an embodiment of the present disclosure, for a potential fault point, if the potential fault point is in a key part of the network-rail-tunnel system, such as the track near a transportation hub, the catenary of the main power supply line, etc., once a fault occurs, it may have a greater impact on the operation of the entire system, and the corresponding severity will also be higher. For example, a potential fault at the turnout of a busy railway main line may cause delays or suspensions of multiple trains. For the fault type, different fault types have different degrees of harm. For example, a fault such as catenary wire breakage has a much higher severity compared to a minor catenary contact fault, because it may directly cause the train to be unable to obtain electrical energy, resulting in large-scale train suspensions. A serious track deformation fault poses a greater threat to the train operation safety than a general track wear fault. For the estimated occurrence time of the potential fault, if the potential fault is expected to occur during the peak train operation period, then its influence range may be wider, because more trains are involved at this time, and the disruption to the transportation order is greater, and the corresponding severity will also increase. On the contrary, if it is expected to occur during the off-peak train operation period, the influence range is relatively small, and the severity may also be lower.

[0127] According to the embodiments of the present disclosure, for potential failures with a high severity level, a higher processing priority is determined. Since the threat to the catenary-track-tunnel system and train operation safety is greater, processing should be arranged as soon as possible to avoid serious consequences. For example, for potential failures of severe track deformation that may cause train derailment, priority must be given to the treatment. Potential failures with a wide influence range also need to be processed with priority to reduce interference to the entire transportation system. For instance, for catenary failures that may affect the operation of trains on multiple lines, maintenance personnel and resources should be arranged for processing with priority.

[0128] According to the determined processing priorities, corresponding processing strategies are formulated. For potential failures with a high priority, a professional repair team needs to be organized immediately for emergency repair, and sufficient repair equipment and spare parts are allocated. For potential failures with a lower priority, processing can be arranged at an appropriate time (such as during train outage periods), or temporary monitoring measures can be taken to monitor their development and then decide on subsequent processing methods. For example, for some potential failures of minor track wear, monitoring can be strengthened first, the wear condition can be checked regularly, and processing can be carried out when the wear reaches a certain level.

[0129] According to the embodiments of the present disclosure, after the failure is repaired, detection is carried out again to confirm that the failure has been completely eliminated. A comprehensive summary of the failure cause, processing method, dispatching effect, etc. of the current failure handling process is made and stored in the case library.

[0130] Figure 4 A block diagram of a failure determination device based on a catenary-track-tunnel system according to an embodiment of the present disclosure is schematically shown.

[0131] As Figure 4 shown, the failure determination device 400 includes a first acquisition module 410, a first determination module 420, and an output module 430.

[0132] The first acquisition module 410 is configured to acquire first real-time operation status information of a target train in a target time period and first real-time operation environment information of the line on which the train travels. Among them, the first real-time operation environment information includes real-time pantograph-catenary monitoring data, real-time track monitoring data, and real-time tunnel monitoring data collected by train monitoring equipment in the same space-time as the real-time operation status information;

[0133] The first determination module 420 is configured to determine a target association relationship among the real-time operation status information, the real-time pantograph-catenary monitoring data, the real-time track monitoring data, and the real-time tunnel monitoring data based on a predetermined map. Among them, the predetermined map is constructed based on historical pantograph-catenary data, historical track data, and historical tunnel data and is used to represent the association relationship among the historical pantograph-catenary data, the historical track data, and the historical tunnel data;

[0134] An output module 430 is configured to input the target association relationship into a trained target graph neural network model, and output potential fault information related to the operating environment of the target train, where the potential fault information includes at least a potential fault point and a fault occurrence probability, so as to perform fault troubleshooting operations on the potential fault point before the target train travels to the potential fault point.

[0135] According to an embodiment of the present disclosure, a method for constructing a predetermined graph includes a second acquisition module, a first construction module, a second determination module, a generation module, a calculation module, and a second construction module.

[0136] The second acquisition module is configured to acquire historical operating state information and historical operating environment information of the target train within a historical time period, where the historical operating state information includes historical driving stability data, and the real-time operating environment information includes historical pantograph-catenary data, historical track data, and historical tunnel data collected by train monitoring devices at the same time and space as the historical operating state information;

[0137] The first construction module is configured to construct at least one transaction dataset according to L data types included in the historical driving stability data, historical pantograph-catenary data, historical track data, and historical tunnel data, where at least one transaction dataset includes L transactions corresponding to the L data types;

[0138] The second determination module is configured to determine a plurality of frequent item sets according to the support degree of each transaction and a support degree threshold;

[0139] The generation module is configured to generate a plurality of frequent sub-item sets corresponding to the frequent item sets according to each transaction in each frequent item set;

[0140] The calculation module is configured to calculate the lift of each frequent sub-item set, and determine the association relationship of each transaction in the frequent sub-item set according to the lift;

[0141] The second construction module is configured to construct a predetermined graph of each transaction according to the association relationship of each transaction.

[0142] According to an embodiment of the present disclosure, the L data types include:

[0143] The historical driving stability data includes at least one of the following data types or a combination of multiple data types: smooth detection data of the target train, instability detection data, vibration detection data, wheel spin and skid data;

[0144] The historical pantograph-catenary data includes at least one of the following data types or a combination of multiple data types: network voltage data of the target train, network current data, pantograph-catenary contact pressure data, arcing rate data of the line, pantograph-catenary foreign object data;

[0145] Historical track data includes at least one of the following data types or a combination of multiple data types: rail profile data, rail corrugation data, turnout abnormal state data, gauge detection data, rail surface state data during wheel spin and skid;

[0146] Historical tunnel data includes at least one of the following data types or a combination of multiple data types: tunnel foreign object intrusion data, tunnel clearance data, tunnel video anomaly data.

[0147] According to an embodiment of the present disclosure, the target graph neural network model is trained by the following method, including a third acquisition module and an optimization module.

[0148] The third acquisition module is configured to acquire a predetermined graph spectrum and historical fault diagnosis information within a historical time period;

[0149] The optimization module is configured to use the predetermined graph spectrum as historical sample data and the historical fault diagnosis information as labels to optimize the initial graph neural network model to obtain the target graph neural network model.

[0150] According to an embodiment of the present disclosure, the fault determination method based on the network-rail-tunnel system further includes a third determination module.

[0151] The third determination module is configured to determine a target fault point from potential fault points according to the second real-time operation state information, the second real-time operation environment information, and the fault occurrence probability of multiple reference trains in the same area, so as to perform fault troubleshooting on the target fault point in advance.

[0152] According to an embodiment of the present disclosure, the potential fault information further includes a potential fault type and a potential fault expected occurrence time. The fault determination method based on the network-rail-tunnel system further includes: a fourth determination module and a fifth determination module.

[0153] The fourth determination module is configured to determine the severity and influence range of the potential fault according to the potential fault point, the potential fault type, and the potential fault expected occurrence time;

[0154] The fifth determination module is configured to determine the processing priority and processing strategy of the potential fault according to the severity and influence range.

[0155] Any one or more of the modules, sub-modules, units, and sub-units according to the embodiments of the present disclosure, or at least part of the functions of any one or more of them, can be implemented in one module. Any one or more of the modules, sub-modules, units, and sub-units according to the embodiments of the present disclosure can be split into multiple modules for implementation. Any one or more of the modules, sub-modules, units, and sub-units according to the embodiments of the present disclosure can be at least partially implemented as a hardware circuit, such as a field programmable gate array (FPGA), a programmable logic array (PLA), a system on a chip, a system on a substrate, a system on a package, an application specific integrated circuit (ASIC), or can be implemented by any other reasonable way of integrating or packaging circuits, etc., in hardware or firmware, or implemented in any one of the three implementation manners of software, hardware, and firmware, or in a suitable combination of any several of them. Alternatively, one or more of the modules, sub-modules, units, and sub-units according to the embodiments of the present disclosure can be at least partially implemented as a computer program module, and when the computer program module is run, the corresponding functions can be executed.

[0156] For example, any one or more of the first acquisition module 410, the first determination module 420, and the output module 430 can be combined and implemented in one module / unit / sub-unit, or any one of the module / unit / sub-unit can be split into multiple modules / units / sub-units. Alternatively, at least part of the functions of one or more of these modules / units / sub-units can be combined with at least part of the functions of other modules / units / sub-units and implemented in one module / unit / sub-unit. According to the embodiments of the present disclosure, at least one of the first acquisition module 410, the first determination module 420, and the output module 430 can be at least partially implemented as a hardware circuit, such as a field programmable gate array (FPGA), a programmable logic array (PLA), a system on a chip, a system on a substrate, a system on a package, an application specific integrated circuit (ASIC), or can be implemented by any other reasonable way of integrating or packaging circuits, etc., in hardware or firmware, or implemented in any one of the three implementation manners of software, hardware, and firmware, or in a suitable combination of any several of them. Alternatively, at least one of the first acquisition module 410, the first determination module 420, and the output module 430 can be at least partially implemented as a computer program module, and when the computer program module is run, the corresponding functions can be executed.

[0157] Figure 5 A block diagram of an electronic device suitable for implementing the method described above according to the embodiments of the present disclosure is schematically shown. Figure 5 The shown electronic device is only an example and should not bring any limitation to the functions and the scope of use of the embodiments of the present disclosure.

[0158] Such as Figure 5As shown, the electronic device 500 according to an embodiment of the present disclosure includes a processor 501, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 502 or a program loaded from a storage section 508 into a random access memory (RAM) 503. The processor 501 can include, for example, a general-purpose microprocessor (such as a CPU), an instruction set processor, and / or a related chipset, and / or a dedicated microprocessor (such as an application-specific integrated circuit (ASIC)), and so on. The processor 501 can also include on-board memory for caching purposes. The processor 501 can include a single processing unit or multiple processing units for performing different actions of the method flow according to an embodiment of the present disclosure.

[0159] In the RAM 503, various programs and data required for the operation of the electronic device 500 are stored. The processor 501, the ROM 502, and the RAM 503 are connected to each other via a bus 504. The processor 501 performs various operations of the method flow according to an embodiment of the present disclosure by executing the programs in the ROM 502 and / or the RAM 503. It should be noted that the program can also be stored in one or more memories other than the ROM 502 and the RAM 503. The processor 501 can also perform various operations of the method flow according to an embodiment of the present disclosure by executing the programs stored in the one or more memories.

[0160] According to an embodiment of the present disclosure, the electronic device 500 can further include an input / output (I / O) interface 505, and the input / output (I / O) interface 505 is also connected to the bus 504. The electronic device 500 can further include one or more of the following components connected to the input / output (I / O) interface 505: an input section 506 including a keyboard, a mouse, etc.; an output section 507 including a cathode ray tube (CRT), a liquid crystal display (LCD), etc., and a speaker, etc.; a storage section 508 including a hard disk, etc.; and a communication section 509 including a network interface card such as a LAN card, a modem, etc. The communication section 509 performs communication processing via a network such as the Internet. A drive 510 is also connected to the input / output (I / O) interface 505 as needed. A removable medium 511, such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, etc., is installed on the drive 510 as needed so that a computer program read from it can be installed into the storage section 508 as needed.

[0161] According to an embodiment of the present disclosure, the method flow according to the embodiment of the present disclosure can be implemented as a computer software program. For example, an embodiment of the present disclosure includes a computer program product, which includes a computer program carried on a computer-readable storage medium, and the computer program includes program code for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from the network through the communication part 509, and / or installed from the removable medium 511. When the computer program is executed by the processor 501, the above functions defined in the system of the embodiment of the present disclosure are executed. According to an embodiment of the present disclosure, the above-described systems, devices, apparatuses, modules, units, etc. can be implemented by computer program modules.

[0162] The present disclosure also provides a computer-readable storage medium, which may be included in the device / device / system described in the above embodiment; or may exist alone without being assembled into the device / device / system. The above computer-readable storage medium carries one or more programs, and when the above one or more programs are executed, the method according to the embodiment of the present disclosure is implemented.

[0163] According to an embodiment of the present disclosure, the computer-readable storage medium may be a non-volatile computer-readable storage medium. For example, it may include but is not limited to: portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination of the above. In the present disclosure, the computer-readable storage medium may be any tangible medium that contains or stores a program, and the program can be used by or in combination with an instruction execution system, device, or device.

[0164] For example, according to an embodiment of the present disclosure, the computer-readable storage medium may include the above-described ROM 502 and / or RAM 503 and / or one or more memories other than ROM 502 and RAM 503.

[0165] An embodiment of the present disclosure also includes a computer program product, which includes a computer program, and the computer program includes program code for executing the method provided by the embodiment of the present disclosure. When the computer program product runs on an electronic device, the program code is used to cause the electronic device to implement the fault determination method based on the network-rail-tunnel system provided by the embodiment of the present disclosure.

[0166] When the computer program is executed by the processor 501, the above functions defined in the system / device of the embodiment of the present disclosure are executed. According to an embodiment of the present disclosure, the above-described systems, devices, modules, units, etc. can be implemented by computer program modules.

[0167] In one embodiment, the computer program may rely on tangible storage media such as optical storage devices and magnetic storage devices. In another embodiment, the computer program may also be transmitted and distributed in the form of a signal on a network medium, and downloaded and installed through the communication part 509, and / or installed from the removable medium 511. The program code included in the computer program can be transmitted using any suitable network medium, including but not limited to: wireless, wired, etc., or any suitable combination of the above.

[0168] According to an embodiment of the present disclosure, the program code for executing the computer program provided by the embodiments of the present disclosure can be written in any combination of one or more programming languages. Specifically, these computing programs can be implemented using high-level procedural and / or object-oriented programming languages, and / or assembly / machine languages. Programming languages include but are not limited to, such as Java, C++, Python, the "C" language, or similar programming languages. The program code can be executed entirely on the user's computing device, partially on the user's device, partially on a remote computing device, or entirely on a remote computing device or server. In cases involving a remote computing device, the remote computing device can be connected to the user's computing device through any type of network, including a local area network (LAN) or a wide area network (WAN), or can be connected to an external computing device (for example, by using an Internet service provider to connect through the Internet).

[0169] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of the present disclosure. In this regard, each block in the flowchart or block diagram may represent a module, a segment of a program, or a portion of code, which contains one or more executable instructions for implementing the specified logical function. It should also be noted that in some alternative implementations, the functions noted in the blocks may occur in a different order than noted in the accompanying drawings. For example, two consecutive blocks shown may actually be executed substantially in parallel, or they may sometimes be executed in the reverse order, depending on the functions involved. It should also be noted that each block in the block diagram or flowchart, and combinations of blocks in the block diagram or flowchart, can be implemented by a dedicated hardware-based system that performs the specified functions or operations, or can be implemented by a combination of dedicated hardware and computer instructions. Those skilled in the art will understand that the features described in the various embodiments of the present disclosure can be combined and / or combined in various ways, even if such combinations or combinations are not explicitly described in the present disclosure. In particular, without departing from the spirit and teachings of the present disclosure, the features described in the various embodiments of the present disclosure can be combined and / or combined in various ways. All such combinations and / or combinations fall within the scope of the present disclosure.

[0170] The embodiments of the present disclosure have been described above. However, these embodiments are for illustrative purposes only and are not intended to limit the scope of the present disclosure. Although the embodiments have been described separately above, this does not mean that the measures in the various embodiments cannot be used advantageously in combination. Without departing from the scope of the present disclosure, those skilled in the art can make various substitutions and modifications, and all such substitutions and modifications should fall within the scope of the present disclosure.

Claims

1. A fault determination method based on a network rail tunnel system, comprising: Acquire first real-time operating status information of the target train in a target time period and first real-time operating environment information of the line on which the target train is traveling, wherein the first real-time operating environment information includes real-time pantograph monitoring data, real-time track monitoring data, and real-time tunnel monitoring data collected by the train monitoring equipment in the same time and space as the real-time operating status information; Based on a predetermined graph, determining a target association relationship between the real-time operation status information, the real-time bow-net monitoring data, the real-time track monitoring data and the real-time tunnel monitoring data, wherein the predetermined graph is constructed based on historical bow-net data, historical track data and historical tunnel data, and is used to characterize the association relationship between the historical bow-net data, the historical track data and the historical tunnel data; The target association relationship is input into a trained target graph neural network model, and potential fault information related to the operating environment of the target train is output, wherein the potential fault information at least includes a potential fault point and a probability of fault occurrence, so that troubleshooting can be performed on the potential fault point before the target train travels to the potential fault point.

2. The method according to claim 1, wherein: The predetermined map is constructed by the following method: Acquire historical operation status information and historical operation environment information of the target train in a historical time period, wherein the historical operation status information includes historical driving stability data, and the real-time operation environment information includes historical pantograph data, historical track data, and historical tunnel data collected by the train monitoring device in the same time and space as the historical operation status information; Constructing at least one transaction data set according to the L data types included in the historical driving stability data, the historical pantograph data, the historical track data, and the historical tunnel data, wherein the at least one transaction data set includes L transactions corresponding to the L data types; According to the support and support threshold of each transaction, multiple frequent item sets are determined; According to each transaction in each frequent item set, a plurality of frequent sub-item sets corresponding to the frequent item sets are generated; Calculating the lift of each frequent sub-item set, and determining the association relationship between each transaction in the frequent sub-item set according to the lift; According to the association relationship between the various transactions, a predetermined graph of each transaction is constructed.

3. The method according to claim 1, wherein: The L data types include: The historical driving stability data includes at least one of the following data types or a combination of multiple data types: stability detection data, instability detection data, vibration detection data, and idling slide data of the target train; The historical pantograph-catwalk data includes at least one of the following data types or a combination of multiple data types: network voltage data, network current data, pantograph-catwalk contact pressure data, line arcing rate data, and pantograph-catwalk foreign matter data of the target train; The historical track data includes at least one of the following data types or a combination of multiple data types: rail profile data, rail corrugation data, turnout abnormal state data, track gauge detection data, and track surface state data during idling and sliding; The historical tunnel data includes at least one of the following data types or a combination of multiple data types: tunnel foreign object intrusion data, tunnel limit data, and tunnel video abnormality data.

4. The method according to claim 1, wherein: The target graph neural network model is trained by the following method: Acquire the predetermined graph and historical fault diagnosis information within the historical time period; The predetermined graph spectrum is used as historical sample data, the historical fault diagnosis information is used as a label, and the initial graph neural network model is optimized to obtain the target graph neural network model.

5. The method according to claim 1, wherein: The method further comprises: According to the second real-time operating status information, the second real-time operating environment information and the fault occurrence probability of multiple reference trains in the same area, a target fault point is determined from the potential fault points so as to troubleshoot the target fault point in advance.

6. The method according to claim 1, wherein: The potential fault information also includes the potential fault type and the estimated time of occurrence of the potential fault. The method also includes: Determine the severity and impact scope of the potential fault according to the potential fault point, the potential fault type and the expected occurrence time of the potential fault; According to the severity and the impact scope, a processing priority and a processing strategy for the potential fault are determined.

7. A fault determination device based on a network rail tunnel system, comprising: A first acquisition module is used to acquire first real-time operating status information of a target train in a target time period and first real-time operating environment information of the route on which the target train is traveling, wherein the first real-time operating environment information includes real-time pantograph monitoring data, real-time track monitoring data, and real-time tunnel monitoring data collected by a train monitoring device in the same time and space as the real-time operating status information; A first determination module is used to determine the target association relationship between the real-time operation status information, the real-time bow-net monitoring data, the real-time track monitoring data and the real-time tunnel monitoring data based on a predetermined map, wherein the predetermined map is constructed based on historical bow-net data, historical track data and historical tunnel data, and is used to characterize the association relationship between the historical bow-net data, the historical track data and the historical tunnel data; An output module is used to input the target association relationship into a trained target graph neural network model, and output potential fault information related to the operating environment of the target train, wherein the potential fault information at least includes a potential fault point and a probability of fault occurrence, so as to perform troubleshooting operations on the potential fault point before the target train travels to the potential fault point.

8. A train comprising: vehicle body; a braking member and a fault determination device based on a network rail tunnel system as described in claim 7.

9. An electronic device, comprising: one or more processors; a memory for storing one or more programs, When the one or more programs are executed by the one or more processors, the one or more processors implement the method according to any one of claims 1 to 6.

10. A computer-readable storage medium having executable instructions stored thereon, which, when executed by a processor, enables the processor to implement the method according to any one of claims 1 to 6.

Images