Rail defect positioning method and device and electronic equipment

Abstract

This application provides a method, apparatus, and electronic device for locating track defects, relating to the field of track detection and analysis technology. The method involves acquiring a dynamic simulation trajectory on a target track section. This dynamic simulation trajectory is generated through bidirectional coupling simulation of the vehicle's trajectory data and a digital twin of the target track section, including a time-varying energy distribution spectrum of stress waves propagating along the track. Based on the time-varying energy distribution spectrum, a track stress wave propagation sequence for the target track section is extracted. This track stress wave propagation sequence includes the measured energy attenuation rate of the track stress waves. An inversion calculation is performed based on the track stress wave propagation sequence to obtain the theoretical energy attenuation rate corresponding to candidate defect points on the target track section. The consistency between the theoretical energy attenuation rate and the measured energy attenuation rate is verified. Candidate defect points that pass the consistency verification are used as the defect location results for the target track section, thereby improving the accuracy and reliability of track defect location.

Description

Technical Field

[0001] This application relates to the field of track inspection and analysis, and in particular to a method, device and electronic equipment for locating track defects. Background Technology

[0002] As railway transportation develops towards high speed and heavy load, the coupled response of railway freight cars to track structure and vehicle system during long-distance operation becomes increasingly complex. Especially in special environments such as mountain tunnels, sleeper settlement, and temperature and humidity fluctuations, the concealment and destructiveness of track structure defects are significantly enhanced.

[0003] Traditional track inspection methods rely heavily on periodic inspections and manual identification, which have problems such as long inspection cycles, large response delays, and limited risk identification dimensions, making it difficult to meet the real-time status perception requirements for critical lines and heavy-load train formations.

[0004] In recent years, some studies have attempted to use stress wave nondestructive testing methods for track defect identification, but they generally suffer from insufficient defect location accuracy, rendering subsequent risk analysis meaningless. This leads to a lack of scientific basis for maintenance decisions and an inability to achieve efficient resource allocation.

[0005] Therefore, how to achieve high-precision positioning of track defects is an urgent problem to be solved. Summary of the Invention

[0006] This application provides a method, apparatus, and electronic device for locating track defects, which can improve the accuracy and reliability of track defect location.

[0007] In a first aspect, embodiments of this application provide a method comprising:

[0008] The dynamic simulation trajectory on the target track section is obtained. The dynamic simulation trajectory is generated by bidirectional coupling simulation based on the vehicle's trajectory data on the target track section and the digital twin of the target track section. The dynamic simulation trajectory includes the time-varying energy distribution spectrum of stress wave propagation along the track.

[0009] The orbital stress wave propagation sequence of the target orbital segment is extracted based on the time-varying energy distribution spectrum. The orbital stress wave propagation sequence includes the measured energy attenuation rate of the orbital stress wave.

[0010] Inversion calculations were performed based on the orbital stress wave propagation sequence to obtain the theoretical energy attenuation rate corresponding to candidate defect points on the target orbital segment.

[0011] The consistency between the theoretical energy decay rate and the measured energy decay rate is verified.

[0012] Candidate defect points that pass the consistency check are used as the defect location results for the target track segment.

[0013] In one possible implementation, inversion calculations are performed based on the orbital stress wave propagation sequence to obtain the theoretical energy attenuation rate corresponding to candidate defect points on the target orbital segment, including:

[0014] Based on the orbital stress wave propagation sequence, the reference characteristic frequencies of the orbital stress waves corresponding to two adjacent reference points on the target orbital section are obtained; the candidate defect points are located on the track between two adjacent reference points, and the reference points are virtual sensor nodes deployed on the target orbital section.

[0015] Based on the relationship function between wavenumber and frequency, the wave value at the reference characteristic frequency is obtained; the relationship function corresponds to the rail type of the target track section.

[0016] Based on the wave value and the reference characteristic frequency, the actual phase velocity corresponding to the reference characteristic frequency is calculated.

[0017] The time difference of arrival of wave peaks between the reference characteristic frequencies corresponding to two adjacent reference points is obtained based on the orbital stress wave propagation sequence.

[0018] The defect source distance of the candidate defect point is calculated based on the actual phase velocity and the time difference of arrival of the wave crest using the differential frequency propagation positioning equation. The defect source distance is the distance between the candidate defect point and the reference point.

[0019] Based on the distance to the defect source and the energy attenuation factor of the target orbital segment, the theoretical energy attenuation rate corresponding to the candidate defect point is calculated.

[0020] In one possible implementation, the consistency verification between the theoretical energy decay rate and the measured energy decay rate includes:

[0021] Based on the orbital stress wave propagation sequence, the measured energy attenuation rate between two reference points corresponding to the candidate defect point is extracted. The reference point is a virtual sensor node on the target orbital segment adjacent to the candidate defect point.

[0022] Obtain the decay rate error between the measured energy decay rate and the theoretical energy decay rate;

[0023] The consistency verification result is obtained based on the absolute value of the attenuation rate error and the error threshold.

[0024] In one possible implementation, obtaining the dynamic simulation trajectory on the target track segment includes:

[0025] Based on the position and attitude information included in the trajectory data, the load of the vehicle on the target track segment is obtained.

[0026] Retrieve the digital twin of the target track segment from the database;

[0027] The load is applied to the digital twin of the track to obtain the elastic deformation field of the track surface under the load and the initial excitation intensity of the stress wave;

[0028] Based on the elastic deformation field and the three-dimensional point cloud model of the rail included in the digital twin of the track, the geometric state of the rail surface after deformation is obtained.

[0029] The contact envelope between the vehicle's wheel tread and the track is obtained based on the deformed track surface geometry using a contact geometry inversion algorithm.

[0030] Based on the contact envelope and the initial excitation intensity of the stress wave, a dynamic simulation trajectory of the target orbit segment is generated.

[0031] In one possible implementation, the orbital stress wave propagation sequence of the target orbital segment is extracted based on the time-varying energy distribution spectrum, including:

[0032] The stress wave observation spectrum is obtained by using a virtual sensor array based on the time-varying energy distribution spectrum. The virtual sensor array is axially deployed on the target orbital segment.

[0033] Based on the relationship function between the wave number and frequency corresponding to the rail type of the track, the stress wave observation spectrum is dispersion compensated to obtain the compensated stress wave spectrum.

[0034] The propagation characteristic parameters are obtained by extracting features from the compensated stress wave spectrum.

[0035] The orbital stress wave propagation sequence of the target orbital segment is formed based on the propagation characteristic parameters.

[0036] In one possible implementation, the vehicle vibration mode set is extracted based on the dynamic simulation trajectory;

[0037] By performing resonance matching on the vehicle body vibration mode set and the track stress wave propagation sequence, the defect type of the target defect point corresponding to the defect location result is determined, and the defect type is either resonance or non-resonance.

[0038] By using the inverse relationship between frequency and vehicle speed pre-built, the critical vehicle speed threshold corresponding to the target defect point is obtained;

[0039] Based on the defect source coordinates, defect type, and critical speed threshold of the target defect point, a correlation matrix between defects and risks is generated, which is used for track operation and maintenance decisions.

[0040] In one possible implementation, the dynamic simulation trajectory further includes the contact envelope between the wheel tread and the track section; the vehicle vibration mode set is extracted based on the dynamic simulation trajectory, including:

[0041] Based on the contact envelope, the longitudinal excitation frequency and lateral excitation amplitude of the vehicle body are obtained.

[0042] Using a pre-constructed multibody dynamics model of the vehicle, the phase angle of the first-order torsional vibration mode and the wheel-rail adhesive oscillation entropy are obtained based on the longitudinal excitation main frequency and the lateral excitation amplitude.

[0043] Based on the longitudinal excitation main frequency, the lateral excitation amplitude, the phase angle of the first-order torsional vibration mode, and the wheel-rail adhesive oscillation entropy, the vehicle body vibration mode set is obtained.

[0044] In one possible implementation, the defect type of the target defect point is determined by resonant matching of the vehicle body vibration mode set and the track stress wave propagation sequence, including:

[0045] The torsional vibration dominant frequency of the target defect point in the vehicle body vibration mode set and the stress wave dominant frequency in the track stress wave propagation sequence were extracted.

[0046] If the torsional vibration frequency and the stress wave frequency satisfy the preset resonance condition, the defect type of the target defect point corresponding to the stress wave frequency is determined to be resonance, and the defect type of other target defect points is non-resonance.

[0047] Secondly, embodiments of this application provide a track defect location device, comprising:

[0048] The simulation trajectory unit is used to acquire the dynamic simulation trajectory on the target track section. The dynamic simulation trajectory is generated by bidirectional coupling simulation based on the vehicle's trajectory data on the target track section and the digital twin of the target track section. The dynamic simulation trajectory includes the time-varying energy distribution spectrum of stress wave propagation along the track.

[0049] The feature extraction unit is used to extract the orbital stress wave propagation sequence of the target orbital segment based on the time-varying energy distribution spectrum. The orbital stress wave propagation sequence includes the measured energy attenuation rate of the orbital stress wave.

[0050] The inversion calculation unit is used to perform inversion calculations based on the orbital stress wave propagation sequence to obtain the theoretical energy attenuation rate corresponding to the candidate defect point on the target orbital segment.

[0051] The consistency verification unit is used to verify the consistency between the theoretical energy decay rate and the measured energy decay rate.

[0052] The defect location unit is used to take the candidate defect points that have passed the consistency check as the defect location results of the target track segment.

[0053] Thirdly, embodiments of this application provide an electronic device, including: a memory and a processor;

[0054] The memory stores the instructions that the computer executes;

[0055] The processor executes computer execution instructions stored in memory, causing the processor to perform the first aspect and / or various possible implementations of the first aspect as described above.

[0056] Fourthly, embodiments of this application provide a computer-readable storage medium storing computer-executable instructions, which, when executed by a processor, are used to implement the first aspect and / or various possible implementations of the first aspect.

[0057] Fifthly, embodiments of this application provide a computer program product, including a computer program that, when executed by a processor, implements the first aspect and / or various possible implementations of the first aspect.

[0058] The track defect location method, device, and electronic equipment provided in this application, on the one hand, obtain a dynamic simulation trajectory based on bidirectional coupling simulation of vehicle trajectory and track digital twin, which can better match the actual structural state of the line and improve the accuracy of the basic data used for subsequent defect inversion; on the other hand, by extracting the track stress wave propagation sequence, detecting the energy attenuation physical characteristics as a key feature of track defect identification, and verifying the consistency between the theoretical energy attenuation rate calculated by inversion and the above-mentioned measured attenuation rate, track defect location is achieved. This effectively filters out pseudo signals that, although abnormal in time or frequency, do not match the defect physical model in their energy attenuation mode. Compared with defect identification from a single time or frequency dimension, this can greatly improve the accuracy and reliability of track defect location. Attached Figure Description

[0059] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this application and, together with the description, serve to explain the principles of this application.

[0060] Figure 1 This is a schematic diagram of one implementation environment involved in this application;

[0061] Figure 2 A flowchart illustrating the track defect location method provided in this application;

[0062] Figure 3 A schematic diagram of the track defect location device provided in this application;

[0063] Figure 4 A schematic diagram of the structure of the electronic device provided in this application.

[0064] The accompanying drawings illustrate specific embodiments of this application, which will be described in more detail below. These drawings and descriptions are not intended to limit the scope of the concept in any way, but rather to illustrate the concept of this application to those skilled in the art through reference to particular embodiments. Detailed Implementation

[0065] Exemplary embodiments will now be described in detail, examples of which are illustrated in the accompanying drawings. When the following description relates to the drawings, unless otherwise indicated, the same numbers in different drawings denote the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this application. Rather, they are merely examples of apparatuses and methods consistent with some aspects of this application as detailed in the appended claims.

[0066] Figure 1 This is a schematic diagram of an implementation environment involved in this application. The implementation environment includes an in-vehicle multi-source sensing unit 10, a digital asset library 20, and a server 30. The in-vehicle multi-source sensing unit 10, the digital asset library 20, and the server 30 are connected by wired or wireless means.

[0067] Server 30 is used to acquire trajectory data of the vehicle on the target track section from the on-board multi-source sensing unit 10, and to acquire the track digital twin of the target track section from the digital asset library 20; it is also used to acquire the dynamic simulation trajectory on the target track section, which is generated by bidirectional coupling simulation based on the trajectory data and the track digital twin, and includes the time-varying energy distribution spectrum of stress wave propagation along the track; it extracts the track stress wave propagation sequence of the target track section based on the time-varying energy distribution spectrum, which includes the measured energy attenuation rate of the track stress wave; it performs inversion calculation based on the track stress wave propagation sequence to obtain the theoretical energy attenuation rate corresponding to the candidate defect point on the target track section; it performs consistency verification between the theoretical energy attenuation rate and the measured energy attenuation rate; and it uses the candidate defect point that passes the consistency verification as the defect location result of the target track section.

[0068] It should be noted that, Figure 1 In the implementation environment shown, server 30 can be a standalone server or a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, content delivery networks (CDN), and big data and artificial intelligence platforms. No restrictions are imposed here.

[0069] In existing technologies, track defect identification generally suffers from insufficient defect location accuracy. The track defect location method provided in this application achieves a substantial improvement in defect location accuracy through a three-step progressive logical structure of "dynamic simulation trajectory - stress wave propagation sequence extraction - energy attenuation consistency verification," thus solving the technical bottleneck of high false alarm rate in traditional methods due to reliance on a single feature.

[0070] The technical solution of this application and how the technical solution of this application solves the above-mentioned technical problems are described in detail below with specific embodiments. These specific embodiments can be combined with each other, and the same or similar concepts or processes may not be described again in some embodiments. The embodiments of this application will now be described with reference to the accompanying drawings.

[0071] Figure 2 A flowchart illustrating the track defect location method provided in this application is shown below. Figure 2 As shown, the method includes:

[0072] S201. Obtain the dynamic simulation trajectory on the target track section. The dynamic simulation trajectory is generated by bidirectional coupling simulation based on the vehicle's trajectory data on the target track section and the digital twin of the target track section. The dynamic simulation trajectory includes the time-varying energy distribution spectrum of stress wave propagation along the track.

[0073] The vehicle's trajectory data in the target track section can include various information generated during the vehicle's operation. This information is collected by the sensing devices included in the onboard multi-source sensing unit, such as the positioning sequence fused from millimeter-wave radar and GNSS (Global Navigation Satellite System), the six-degree-of-freedom attitude angles output by the inertial measurement unit, and the cumulative displacement recorded by the wheel and axle encoders.

[0074] Track digital twins are stored in a digital asset repository and possess a unified data numbering system and metadata tags. Track digital twins corresponding to track sections are typically obtained by filtering using the track section's number (e.g., line ID + start and end mileage). They can also be obtained through timestamp filtering or API calls, such as specifying the data collection date to retrieve the corresponding track digital twin, or by calling the API (Application Programming Interface) provided by the engineering section's asset system. Furthermore, the track digital twins stored in the digital asset repository have standardized data formats, supporting multiple format conversions to adapt to track digital twin modeling tools.

[0075] In this embodiment, a high-fidelity mapping of the physical behavior of the track structure is achieved by constructing a dynamic simulation trajectory. The dynamic simulation trajectory is constructed through bidirectional coupling simulation based on the vehicle's actual trajectory data and a digital twin of the track that reflects the physical properties of the target track section. This accurately recreates the propagation path and energy distribution characteristics of stress waves within the track, more closely reflecting the actual structural state of the line and improving the data fidelity of defect inversion from the source. Furthermore, the subsequent defect location steps are all performed within the simulation environment provided by the track digital twin, avoiding interference from other noise data. Therefore, precise defect location of the target track section can be achieved based on the dynamic simulation trajectory.

[0076] S202. Extract the orbital stress wave propagation sequence of the target orbital segment based on the time-varying energy distribution spectrum. The orbital stress wave propagation sequence includes the measured energy attenuation rate of the orbital stress wave.

[0077] In this embodiment, the dynamic simulation trajectory includes the time-varying energy distribution spectrum of the stress wave propagating along the track. By extracting features from the time-varying energy distribution spectrum, various parameters generated when the stress wave propagates on the target track segment can be obtained, such as the measured energy attenuation rate of the track stress wave.

[0078] S203. Based on the propagation sequence of orbital stress waves, perform inversion calculations to obtain the theoretical energy attenuation rate corresponding to the candidate defect points on the target orbital section.

[0079] This embodiment uses the parameters generated when the stress wave propagates in the target track section, which are included in the track stress wave propagation sequence, to invert and calculate the theoretical energy attenuation rate corresponding to the candidate defect point on the target track section. This ensures that each candidate point has a clear corresponding theoretical energy attenuation rate, providing a physical discrimination benchmark for the defect point on the target track section.

[0080] S204. Verify the consistency between the theoretical energy decay rate and the measured energy decay rate.

[0081] S205. The candidate defect points that pass the consistency check are used as the defect location results of the target track section.

[0082] This embodiment elevates defect determination from traditional empirical threshold judgment to scientific verification based on the principle of wave energy conservation. It places candidate defect points under the dual constraints of "theoretical expectation" and "actual observation" to verify their physical authenticity. Specifically, it verifies the consistency between the theoretical energy decay rate and the measured energy decay rate to achieve the essential verification of "whether the candidate defect conforms to the physical laws of wave propagation".

[0083] Consistency checks can identify interference signals that exhibit energy anomalies but whose attenuation patterns do not conform to the theoretical defect model. Examples include abnormal correlations between attenuation rate and propagation distance, or mismatches between attenuation gradient and structural boundary conditions. Only when the measured attenuation characteristics fully conform to the theoretical defect propagation model, i.e., when the consistency check is passed, are the corresponding candidate defect points confirmed as real defects, thus significantly reducing the false alarm rate at the physical mechanism level.

[0084] The track defect location method provided in this application realizes a high-fidelity mapping of the physical behavior of the track structure by constructing a dynamic simulation trajectory, and extracts the track stress wave propagation sequence containing the measured energy attenuation rate from it, thus expanding the defect identification dimension from traditional time or frequency features to the deep physical feature of energy dissipation; after obtaining the theoretical energy attenuation rate based on the propagation sequence, the consistency between the theoretical value and the measured value is verified.

[0085] As can be seen, this application utilizes a three-step progressive logical structure of "dynamic simulation trajectory - stress wave propagation sequence extraction - energy attenuation consistency verification" to achieve a substantial improvement in defect location accuracy. Furthermore, a candidate point is only confirmed as a real defect when its energy attenuation pattern simultaneously conforms to the wave propagation theory and simulation observation results. This effectively filters out pseudo-signals that, although exhibiting anomalies in time or frequency, do not match the physical model of the defect in their energy attenuation pattern. Compared to defect identification from a single time or frequency dimension, this approach can greatly improve the accuracy and reliability of track defect location.

[0086] In an exemplary embodiment of this application, the step of obtaining the theoretical energy attenuation rate corresponding to a candidate defect point on the target orbital segment by performing inversion calculation based on the orbital stress wave propagation sequence may specifically include:

[0087] Based on the orbital stress wave propagation sequence, the reference characteristic frequencies of the orbital stress waves corresponding to two adjacent reference points on the target orbital section are obtained; the candidate defect points are located on the track between two adjacent reference points, and the reference points are virtual sensor nodes deployed on the target orbital section.

[0088] Based on the relationship function between wavenumber and frequency, the wave value at the reference characteristic frequency is obtained; the relationship function corresponds to the rail type of the target track section.

[0089] Based on the wave value and the reference characteristic frequency, the actual phase velocity corresponding to the reference characteristic frequency is calculated.

[0090] The time difference of arrival of wave peaks between the reference characteristic frequencies corresponding to two adjacent reference points is obtained based on the orbital stress wave propagation sequence.

[0091] The defect source distance of the candidate defect point is calculated based on the actual phase velocity and the time difference of arrival of the wave crest using the differential frequency propagation positioning equation. The defect source distance is the distance between the candidate defect point and the reference point.

[0092] Based on the distance to the defect source and the energy attenuation factor of the target orbital segment, the theoretical energy attenuation rate corresponding to the candidate defect point is calculated.

[0093] The orbital stress wave propagation sequence includes not only the measured energy attenuation rate of the orbital stress wave, but also the energy proportion of the characteristic frequency band corresponding to each node on the target orbital segment and the arrival time difference of the stress wave peaks of adjacent nodes. Each node is a virtual sensor node deployed on the target orbital segment, and the characteristic frequency band is a predefined, fixed risk-sensitive frequency band (e.g., 50-80Hz). The energy proportion of the characteristic frequency band is used to describe the overall energy distribution characteristics of the stress wave.

[0094] In this embodiment, to improve the accuracy and reliability of track defect location by verifying the consistency between theoretical and measured values, the theoretical energy attenuation rate corresponding to the candidate defect point on the target track segment is first obtained. Specifically, two adjacent virtual sensor nodes on both sides of the candidate defect point are used as reference points, and the theoretical energy attenuation rate corresponding to the candidate defect point is calculated based on the defect source distance between the candidate defect point and the reference points.

[0095] To achieve high-precision time-delay localization, the distance between the candidate defect point and the corresponding reference point is calculated. First, the reference characteristic frequency of the reference point is obtained based on the energy proportion of the characteristic frequency band corresponding to each node in the orbital stress wave propagation sequence. The reference characteristic frequency is the dominant frequency component within the characteristic frequency band corresponding to that node. It can be determined by finding the frequency corresponding to the power spectrum peak within this characteristic frequency band, or by weighted averaging of the energy within the characteristic frequency band.

[0096] After obtaining the reference characteristic frequency, the relationship function between the wavenumber and frequency corresponding to the rail type of the target track section is obtained. Based on the relationship function and the reference characteristic frequency, the wave value at the reference characteristic frequency is obtained. Based on the wave value and the reference characteristic frequency, the actual phase velocity corresponding to the reference characteristic frequency is calculated to achieve frequency-varying wave velocity correction. Then, the wave peak arrival time difference between the reference characteristic frequencies corresponding to two reference points is extracted from the track stress wave propagation sequence. Subsequently, the defect source distance of the candidate defect point is calculated based on the actual phase velocity and the wave peak arrival time difference using the difference frequency propagation positioning equation. The defect source distance is the distance between the candidate defect point and the reference point.

[0097] After calculating the time delay of candidate defect points by measuring the defect source distance, the theoretical energy attenuation rate corresponding to the candidate defect point is calculated based on the defect source distance and the energy attenuation factor of the target track section. Therefore, based on the dual consistency criteria of time delay positioning and energy attenuation rate, false alarms or spurious defects caused by structural non-uniformity can be eliminated, improving the spatial accuracy and fault tolerance of defect source positioning. This method is suitable for high-grade railways and tracks in complex terrain. The energy attenuation factor is related to the rail material, frequency, and track structure (including ballast compaction) of the target track section and is a fundamental parameter of the target track section; its acquisition process will not be elaborated here.

[0098] Thus, through the above embodiments, this application utilizes the measured characteristic frequency and wave crest time difference of adjacent reference points, combined with the relationship function between the wave number and frequency specific to the rail type, to dynamically correct the frequency-varying wave velocity. Then, it accurately inverts the defect source distance using the difference-frequency propagation equation, fundamentally overcoming the positioning error of traditional fixed wave velocity models in dispersive media. Furthermore, based on this precise distance and the energy attenuation factor related to the track structure, the theoretical attenuation rate is calculated, allowing the theoretical value to be closely coupled with the spatial location of the defect point and the local track physical state. This provides a high-fidelity, traceable physical benchmark for subsequent consistency verification with measured values. In addition, the entire process forms a closed-loop physical inversion chain of "feature extraction - wave velocity correction - difference-frequency positioning - attenuation calculation," ensuring adaptability and strong robustness under different rail types and complex track conditions, thereby significantly improving the overall accuracy and reliability of defect positioning.

[0099] In an exemplary embodiment of this application, the step of verifying the consistency between the theoretical energy decay rate and the measured energy decay rate may specifically include:

[0100] Based on the orbital stress wave propagation sequence, the measured energy attenuation rate between two reference points corresponding to the candidate defect point is extracted. The reference point is a virtual sensor node on the target orbital segment adjacent to the candidate defect point.

[0101] Obtain the decay rate error between the measured energy decay rate and the theoretical energy decay rate;

[0102] The consistency verification result is obtained based on the absolute value of the attenuation rate error and the error threshold.

[0103] In this embodiment, based on the physical consistency requirement of the principle of conservation of wave energy, the consistency verification between the theoretical energy decay rate and the measured energy decay rate is carried out, and the defect location of the target orbit segment is transformed into a clear and calculable numerical judgment. Thus, at the algorithm level, an objective and unambiguous automated decision-making rule is formed, avoiding the uncertainty and inconsistency caused by subjective judgment based on human experience in traditional methods.

[0104] Specifically, based on the orbital stress wave propagation sequence, the measured energy attenuation rate between two reference points corresponding to the candidate defect point is extracted. This ensures that the consistency verification targets the most direct local energy anomaly that may be caused by the candidate point, effectively eliminating interference from energy changes in distant structures or unrelated sections. This allows the verification results to accurately and sensitively reflect the structural state at that specific location. Then, the attenuation rate error between the measured and theoretical energy attenuation rates is obtained. The absolute value of the attenuation rate error is compared with an error threshold, enabling the distinction between significant energy anomalies exceeding the tolerance caused by actual defects and normal fluctuations within the allowable error threshold caused by background noise, material inhomogeneities, or minor measurement fluctuations, based on the consistency verification results.

[0105] Thus, through the above embodiments, this application implements consistency verification as a standardized comparison process based on local measured data and with clearly quantified thresholds. This not only endows the defect judgment process with a high degree of objectivity and automation, but also ensures that the judgment results can both accurately capture real defect signals and robustly resist various interferences by focusing on local verification and threshold control, ultimately significantly improving the accuracy and engineering applicability of the defect location method provided by this application.

[0106] Based on the above embodiments, in an exemplary embodiment provided in this application, defect location is performed in a three-dimensional waveguide mesh constructed from a track digital twin, based on the peak arrival time difference ΔT extracted from the track stress wave propagation sequence and the measured energy attenuation rate η. This embodiment introduces a rail wavenumber-frequency relationship function to replace the traditional fixed wave velocity assumption, achieving dynamic correction of the frequency-varying wave velocity characteristics of the target track section, and combining it with the track digital twin to construct a model of a three-dimensional waveguide mesh suitable for different track types and structural states.

[0107] The track digital twin includes a 3D point cloud model of the rails, a ballast compaction gradient map, and a temperature-humidity coupled Young's modulus correction coefficient table. The 3D point cloud model of the rails includes rail gap coordinate annotations. When constructing the 3D waveguide mesh corresponding to the target track section, the 3D point cloud model of the rails provides the geometric skeleton and rail gap location, the ballast compaction gradient map provides the boundary damping and stiffness distribution, and the Young's modulus correction coefficient table provides the local material properties (Young's modulus). The 3D waveguide mesh is a fusion and discretization representation of the various data contained in the track digital twin. It serves as the computational carrier for stress wave propagation simulation and the physical basis for defect inversion and localization.

[0108] The steps for performing defect localization in a 3D waveguide mesh include:

[0109] Frequency-dependent wave speed correction: Based on the rail type, look up the frequency from the wave number-frequency relationship function. Corresponding wave number Calculate the actual phase velocity: ;

[0110] in, Represents frequency The actual phase velocity of the stress wave Indicates the wave number corresponding to the frequency. This represents the characteristic frequency of the stress wave.

[0111] Time-delay localization equation: Using the waveforms of two adjacent virtual sensor nodes of a candidate defect point as a reference, the distance to the defect source is calculated using the difference frequency propagation principle. (Distance between the candidate defect point and the first sensor):

[0112] ;

[0113] in, This represents the characteristic frequencies corresponding to two virtual sensor nodes. This represents the time difference between the arrival times of the wave peaks at two frequencies. This represents the sleeper reflection correction amount, used to eliminate multipath errors caused by sleepers. This indicates the location distance of the defect source relative to the reference node.

[0114] The virtual sensor nodes are typically deployed at the sleeper locations on the target track section. The sleeper reflection correction is defined as follows: ;in, For the sleeper inclination angle, As the dominant frequency, This represents the phase velocity at the corresponding frequency. Here, f is usually taken as a representative dominant frequency within the current analysis segment. It can be either f1 or f2 (e.g., the one with higher energy), or some average value between f1 and f2 (e.g., energy-weighted average), or another dominant frequency extracted from the signal segment.

[0115] Energy decay consistency verification: Verify the measured energy decay rate Is the error between the model and the theoretical model within the tolerance range? ;in, As the energy decay factor, This represents the distance to the defect source.

[0116] The final output is the spatial coordinates of the defect source that satisfy the dual consistency constraints of time delay and energy, including: latitude and longitude coordinates; track gauge chain number (track position code).

[0117] This embodiment utilizes the propagation characteristics of stress waves, combined with time difference and energy attenuation information, to accurately determine the spatial location of track defects. The track defect localization method implemented in this embodiment consists of three core steps, each calculated based on a physical model and track structural characteristics, ensuring that the localization results possess both temporal accuracy and energy consistency.

[0118] I. Frequency-Velocity Correction: Since the target track section, being a rail, is a waveguide with a complex geometry, stress waves of different frequencies propagate at different speeds (i.e., dispersion effect). This step calculates the actual phase velocity by looking up the corresponding wavenumber from a pre-established wavenumber-frequency relationship table based on the rail type, instead of using a fixed wave velocity value. This avoids wave velocity misjudgment when there are rail bends, turnouts, or temperature and humidity changes, thereby improving the timing accuracy of subsequent positioning.

[0119] II. Time Delay Positioning Calculation: Two adjacent virtual sensors are selected, and the time difference between the wave peak signals received by them at two different frequencies is measured. Using the phase velocity difference corresponding to the aforementioned frequencies, the positioning formula based on the difference frequency propagation principle is substituted to calculate the spatial distance between the defect source and the first sensor. In order to avoid multipath interference caused by the sleeper structure to wave propagation, a sleeper reflection correction is also introduced to compensate for the distance value, thereby enhancing the positioning stability in dense sleeper areas.

[0120] III. Energy Attenuation Consistency Verification: After completing the time delay location, it is necessary to verify whether the theoretical energy attenuation value at that location is consistent with the measured value. This step compares the actually measured energy attenuation rate with the theoretical attenuation value derived from the path distance and ballast compaction, requiring that the error between the two be within the allowable range. If they are inconsistent, it indicates that the point may be a false defect or a misjudged point, and it will be eliminated. Only points that simultaneously satisfy the time positioning equation and energy consistency verification are confirmed as the true defect source location. The final output defect coordinates include its latitude and longitude and track gauge chain number, which can be directly used in the maintenance dispatch system or speed limit warning system.

[0121] Thus, through the above embodiments, this application establishes a triple mechanism of "frequency-varying wave velocity correction + time delay calculation + energy consistency," achieving high-precision spatial positioning of track defect sources. This avoids the misjudgment problems caused by incorrect wave velocity assumptions or structural interference in traditional single time difference methods, and is an intelligent inversion method driven by enhanced physical constraints and structural perception. Compared to traditional schemes that rely solely on time difference positioning and are susceptible to false defects due to wave velocity assumption errors and environmental noise, the track defect positioning method provided in this application, by introducing energy attenuation consistency verification, can effectively identify and eliminate signal anomalies caused by non-defect factors. Only when both time delay and energy characteristics satisfy the physical propagation laws is it determined to be a real defect.

[0122] In an exemplary embodiment of this application, the step of obtaining the dynamic simulation trajectory on the target track segment may specifically include:

[0123] Based on the position and attitude information included in the trajectory data, the load of the vehicle on the target track segment is obtained.

[0124] Retrieve the digital twin of the target track segment from the database;

[0125] The load is applied to the digital twin of the track to obtain the elastic deformation field of the track surface under the load and the initial excitation intensity of the stress wave;

[0126] Based on the elastic deformation field and the three-dimensional point cloud model of the rail included in the digital twin of the track, the geometric state of the rail surface after deformation is obtained.

[0127] The contact envelope between the vehicle's wheel tread and the track is obtained based on the deformed track surface geometry using a contact geometry inversion algorithm.

[0128] Based on the contact envelope and the initial excitation intensity of the stress wave, a dynamic simulation trajectory of the target orbit segment is generated.

[0129] Among them, the dynamic simulation trajectory is generated by bidirectional coupling simulation based on the trajectory data of the vehicle in the target track section and the digital twin of the track in the target track section. The trajectory data of the vehicle in the target track section includes position information and attitude information, such as positioning sequence, six-degree-of-freedom attitude angle and cumulative displacement of travel, etc. The trajectory data are collected by various sensing devices included in the on-board multi-source sensing unit.

[0130] Positioning sequence: generated by fusion of millimeter-wave radar and GNSS, with update frequency satisfying... .

[0131] Six-DOF attitude angles: Output by the Inertial Measurement Unit (IMU), including: ;in, These are roll angle, pitch angle, and yaw angle, respectively. It is a triaxial acceleration.

[0132] Cumulative displacement of travel: recorded by the wheel and axle encoder, forming a displacement closed-loop verification value. ; This represents the cumulative displacement of the travel (unit: meters), representing the distance from the initial moment to the current time. The total distance traveled, Let represent the integral variable, representing the time from the initial time 0 to the current time. At any point in the middle, Indicates at time The instantaneous travel speed is calculated by the wheel axle encoder or IMU.

[0133] Millimeter-wave radar + GNSS positioning sequence constructs the entire vehicle trajectory and binds it to the track digital twin; it is used to align stress wave inversion coordinates with the track coordinate system, ensuring spatial consistency of positioning output; it provides a map projection basis for "defect source coordinates," making the output maintenance coordinates operable; six-degree-of-freedom attitude angles capture micro-tilt, pitch, and other attitude fluctuations during vehicle operation, correcting the normal load distribution at the wheel-rail contact point, which affects tread contact strength and excitation frequency. The cumulative displacement (wheel and axle encoder) can provide displacement closed-loop in GNSS failure scenarios (such as tunnels and canyons), achieving trajectory continuity; it provides a "simulation interpolation" interval for trajectory discontinuities, used to correct the spatiotemporal alignment between positioning points and waveguide inversion coordinates (e.g., a defect point located in a GNSS signal loss segment can still be accurately located).

[0134] The target track section's corresponding digital twin obtained from the database includes a 3D point cloud model of the rail, a ballast density gradient map, and a temperature-humidity coupled Young's modulus correction coefficient table for the rail. The 3D point cloud model of the rail includes rail gap coordinate annotations.

[0135] 3D point cloud model of the rail (including rail gap coordinate annotation): used to establish the realistic rail surface topology and provide boundaries for subsequent contact geometry inversion; Table of Young's modulus correction coefficients for temperature-humidity coupling of the rail: ;

[0136] in, This represents the corrected Young's modulus of the rail. This represents the Young's modulus under reference temperature and humidity conditions. Indicates the current temperature and humidity. Indicates the reference temperature and humidity. This indicates the influence coefficient of temperature and humidity.

[0137] A 3D point cloud model of rails is a reconstruction of the three-dimensional geometry of the rail body and surrounding structure based on high-density spatial data collected by LiDAR, track inspection vehicles, or track inspection robots. Each point in the point cloud data represents the coordinate value of a certain location in space, and the overall model reflects the actual geometric state of the rail (such as gauge, rail head wear, lateral offset, etc.). Rail gaps are the joint areas connecting rail sections, used for thermal expansion and contraction adjustment. The location of rail gaps is identified by abrupt changes in point cloud spacing, and their center coordinates and numbers are manually or automatically marked for reference in subsequent identification of stress wave sources and defect inversion and location. Retrieval method: In the digital asset library of the track maintenance section, which serves as the database, rail point cloud data is usually archived in a standardized coordinate system format and includes metadata, including acquisition time, section number, and offset reference. The point cloud model of the target track section can be retrieved by searching and retrieving the section number through the API interface or query interface.

[0138] The ballast compaction gradient map is a spatially quantified map of the compaction degree of crushed stone (ballast) under the track bed, reflecting the support stiffness gradient of the track structure. High compaction indicates strong ballast stability, while low compaction may lead to track bed subsidence or abnormal structural resonance. The track bed is inspected using ground-penetrating radar (GPR), hammer impact rebound testing, or track bed CT scanning equipment. The compaction data is mapped to track direction and lateral coordinates to generate a two-dimensional or three-dimensional distribution map, where each pixel represents the compaction rate per unit volume in a certain area. Retrieval method: This image data is generally stored in raster form within the track status subsystem of the engineering section, bound to the track section number and collection batch. The compaction distribution map of the target section can be retrieved through a geographic information system (GIS platform) or database field query.

[0139] In this embodiment, after obtaining the vehicle's load on the target track segment based on the position and attitude information included in the trajectory data, the load is applied to the track digital twin corresponding to the target track segment to simulate the vehicle's form in the simulation environment provided by the track digital twin, thereby obtaining the elastic deformation field and initial excitation intensity of the stress wave on the track surface under load; then, based on the elastic deformation field and the three-dimensional point cloud model of the rail included in the track digital twin, the geometric state of the deformed track surface is obtained; through the contact geometry inversion algorithm, the contact envelope between the vehicle's wheel tread and the track is obtained based on the geometric state of the deformed track surface; based on the contact envelope and the initial excitation intensity of the stress wave, the dynamic simulation trajectory of the track is generated.

[0140] Based on the above embodiments, in an exemplary embodiment provided by this application, bidirectional coupling simulation mainly includes forward "load application" and reverse "contact geometry inversion", as well as "dynamic simulation trajectory generation".

[0141] Load application: Applying the axle redistribution load based on trajectory data. Applied to the orbital digital twin, calculations were performed using an explicit dynamics solver:

[0142] Elastic deformation field of the rail surface under wheel load: ;

[0143] Initial excitation intensity of stress wave at the rail gap: The initial value of the stress wave generated by deformation and impact at the rail joint is the starting point for stress wave propagation.

[0144] in, This represents the displacement distribution on the rail surface (elastic deformation field). This indicates the axle load distribution of the truck. Indicates the initial intensity of the stress wave. For contact area, This represents the space-time influence function, describing the rail surface response as a function of position under load. and time The function of the variation law is used to express the dynamic response of external excitation (such as wheel load) propagation and action in the track structure. It is a concept widely used in elasticity and structural dynamics. This represents the equivalent Young's modulus after temperature and humidity correction, derived from the "Temperature and Humidity-Young's Modulus Correction Coefficient Table" in the digital twin, reflecting the influence of the environment on rail stiffness.

[0145] During the load application process, the following methods were used: It is mainly used to model the following two aspects: Spatial influence: the degree of influence of external loads on different surrounding points decreases with distance; Temporal responsiveness: since the load is a moving load that changes with time, the response also has time-varying characteristics (such as delay, reflection, etc.). This function, as a kernel function, modulates the distribution law of elastic deformation of the track surface. Spatial distribution effect ( Dimension): Function Follow The changes reflect the "propagation length" and "local range of influence" of the force propagating on the track surface. They exhibit the following characteristics:

[0146] The peak value is highest near the contact point; the amplitude gradually decreases as it expands outward; slight reverse fluctuations may occur (such as reflections caused by rail gaps or non-uniform structures).

[0147] Time response effect ( (Dimension): Because the load is moving, It describes the relative time delay of the response at different locations, reflecting the "propagation characteristics of fluctuations": the response near the load point appears first; the response far from the load point has a propagation delay; the response on the time axis usually exhibits an impact decay type or an oscillation type.

[0148] In addition, in physical simulation, Determined numerically from structural dynamics equations (such as wave equations or Green's function solutions), it is the response weighting function of the "medium structure" in the load propagation path. It will input the load. Converted into actual displacement response This indirectly affects the subsequent stress wave excitation intensity, spectral characteristics, and defect location results. Spatial-temporal influence function. As a modulation factor, the standard load With the elastic modulus of the rail The resulting reference displacement amplitude is physically modulated by the actual response at different spatial locations and time points. Furthermore, the actual orbit is not a uniform continuum. It can reflect the response heterogeneity caused by local non-uniformity such as ballast voids and loose sleepers.

[0149] Contact geometry inversion: Based on the 3D point cloud model of the rail, and combined with the elastic deformation results under wheel load, the rail surface geometry is dynamically constructed. Specifically, the 3D point cloud model of the rail provides the initial geometry of the rail surface before it is subjected to force, including spatial details such as rail head curvature and rail gap location. In the explicit dynamics solver, the elastic deformation of the rail surface at various locations is calculated based on the load Pi, and the deformed rail surface point set or continuous surface is output as the deformed rail surface geometry.

[0150] Then, based on the deformed rail surface curvature, normal vector, and rail gap boundary, combined with the wheel tread geometry, a contact geometry inversion algorithm is used to calculate the contact envelope between the wheel tread and the rail: ;in, This represents the contact constraint function between the wheel and the rail;

[0151] In the contact geometry inversion process, the explicit dynamics solver is a numerical computation tool used to solve high-speed, transient dynamic problems. It performs response calculations step-by-step and explicitly, under known initial states and load conditions. In this application, the explicit dynamics solver simulates the elastic deformation of the rail surface, stress wave excitation, and propagation process after wheel loads are applied to the rail. Its calculation method does not require solving large-scale equation systems, making it suitable for handling rapidly changing problems such as collisions, contact, and fluctuations. It can achieve rapid response on edge devices or chips, meeting the needs of real-time trajectory simulation and waveguide modeling. In short, the explicit dynamics solver, with high time resolution and low computational latency, drives the track digital twin to respond in real-time to changes in vehicle loads, serving as a key supporting module for realizing dynamic simulation trajectories.

[0152] Dynamic simulation trajectory generation: Considering track vibration feedback, output the simulation trajectory, including:

[0153] Spatiotemporal coordinates of the wheel-rail contact point obtained from the contact envelope: ;

[0154] Time-varying energy distribution spectrum of stress wave propagation along the rail: ;

[0155] in, Indicates the wheel-rail contact point at time [time]. The three-dimensional spatiotemporal coordinate vector, Let be the coordinates of the contact point in the longitudinal direction of the track. Let be the coordinates of the contact point in the horizontal (left-right) direction. Let be the coordinates of the contact point in the vertical direction. For time variables, Indicates the location ,time Stress wave energy intensity at the location, The initial excitation energy of the stress wave (source point energy) The energy attenuation factor during the propagation of stress waves along the rail is affected by the ballast density gradient diagram. These are the position coordinates along the rail direction. The angular frequency of the stress wave reflects the rate of change of the wave over time. Let be the wave number of the stress wave, representing the propagation characteristics of the wave in space.

[0156] When stress waves propagate in the rail, part of their energy is conducted to the sleepers through the rail base. The higher the density of the ballast, the more energy is coupled and absorbed, and the faster the attenuation. Conversely, insufficient density causes the rail to float, which can lead to rebound or local resonance, resulting in local energy stagnation or distorted propagation. Therefore, the density of the ballast directly determines the energy dissipation characteristics of the waveguide system. Values. Energy decay factor. The spatial distribution values ​​can be determined using the following empirical model:

[0157] ;

[0158] in, Indicates the position of the rail The local energy decay factor at that location This represents the base attenuation factor under the reference density (design value). The density gradient influence coefficient needs to be calibrated based on historical tests or actual measurements. Indicates the ballast density gradient at The spatial derivative at a given location is expressed as the rate of change in ballast density per meter. First, a ballast density map containing the track section is retrieved from the database. Each point in the map represents the ballast compaction rate at the current location. Then, these compaction rate values ​​are extracted sequentially along the rail direction to form a one-dimensional density sequence. Finally, the density gradient at each location is calculated by dividing the density change between adjacent points by the corresponding spatial distance, thus obtaining the spatial derivative. Substituting the density gradient value into the aforementioned empirical model, combined with measured or simulated calibration values... and This will output the local data for each track segment. value.

[0159] Thus, through the above embodiments, this application first transforms macroscopic vehicle trajectory data into microscopic wheel-rail contact mechanical response, and accurately calculates the elastic deformation field of the rail surface and the initial excitation intensity of stress waves using a physical solver, thereby providing a real and reliable physical excitation source for subsequent wave analysis. Second, based on the deformed rail surface after load application and a high-precision rail point cloud model, a contact geometry inversion algorithm is used to dynamically reconstruct the instantaneous contact envelope between the wheel and the rail, breaking through the limitations of the traditional rigid contact assumption and accurately characterizing the boundary conditions for stress wave generation. Finally, the generated dynamic simulation trajectory integrates the spatiotemporal coordinates of the contact point and the time-varying energy spectrum of the stress wave, forming a holographic physical snapshot of the coupling effect between the vehicle and the rail, laying a structured and interpretable data foundation for the subsequent extraction of high-fidelity stress wave propagation sequences, and ensuring the physical consistency and analytical reliability of the input data of the entire defect location method from the source.

[0160] In an exemplary embodiment of this application, the step of extracting the orbital stress wave propagation sequence of the target orbital segment based on the time-varying energy distribution spectrum may specifically include:

[0161] The stress wave observation spectrum is obtained by using a virtual sensor array based on the time-varying energy distribution spectrum. The virtual sensor array is axially deployed on the target orbital segment.

[0162] Based on the relationship function between the wave number and frequency corresponding to the rail type of the track, the stress wave observation spectrum is dispersion compensated to obtain the compensated stress wave spectrum.

[0163] The propagation characteristic parameters are obtained by extracting features from the compensated stress wave spectrum.

[0164] The orbital stress wave propagation sequence of the target orbital segment is formed based on the propagation characteristic parameters.

[0165] Rails are non-ideal waveguide structures, exhibiting dispersion effects. In actual operation, rails are not uniform, boundless, ideal media; stress waves of different frequencies propagate at different speeds within them, resulting in dispersion. Without correction, measured signals exhibit waveform broadening and time delay drift, severely impacting defect location accuracy. Currently, unprocessed time-domain signals are directly used for peak identification and propagation time calculation, ignoring the physical law of wavenumber variation with frequency. This leads to location results deviating from the actual structure, rendering the system unusable for critical lines or high-risk sections. The geometry of the rail cross-section determines its dispersion characteristics. Therefore, this embodiment constructs a track stress wave propagation sequence through virtual sensor array deployment, dispersion effect compensation, and quantitative extraction of propagation characteristics.

[0166] Specifically, the virtual sensor array is deployed by setting up multiple virtual sensor nodes at equal intervals along the axial direction of the target track section, with a sampling frequency higher than 1kHz, to ensure the temporal accuracy and spatial coverage of typical stress wave frequency bands (such as 50-80Hz).

[0167] Dispersion effect compensation: The stress wave observation spectrum is obtained based on the time-varying energy distribution spectrum using a virtual sensor array. The observed waveform, characterized by the stress wave observation spectrum at each node, is multiplied by a compensation factor in the frequency domain. This reverses the phase broadening caused by the change in wave velocity with frequency, making the signal waveform closer to the real source wave.

[0168] Propagation Feature Quantization Extraction: Extract the following key features from the compensated waveform: peak arrival time difference Used for inverting the distance to the defect source; characteristic frequency band energy percentage. Used to identify whether structural resonance has been triggered; energy decay rate Used to determine the ballast support status or the scattering effect of rail gaps.

[0169] Constructing the propagation sequence matrix: Integrating the features of all nodes to form a multi-dimensional matrix. This serves as the standard input structure for the orbital stress wave propagation sequence, which is then used by the subsequent positioning engine and risk assessment module.

[0170] Based on the above embodiments, in an exemplary embodiment provided by this application, in order to extract the track stress wave propagation sequence of the target track segment from the dynamic simulation trajectory, the time-varying energy distribution spectrum of the stress wave is first separated from the dynamic simulation trajectory, and a virtual sensor array is deployed axially along the target track segment with a sampling frequency ≥1kHz. That is, the simulation recreates the law of energy variation with time and position during the propagation of the stress wave along the rail, and sampling and monitoring are performed by simulating the deployment of multiple "virtual sensors" along the length of the rail. For each virtual sensor node, the following processing is performed:

[0171] Dispersion effect compensation: through dispersion relation function , for the observed waveform Perform compensation to restore the original waveform: ;

[0172] in, For frequency variables, This is a wavenumber-frequency relationship function, derived from rail type parameters (rail cross-sectional shape). This refers to the relative position of the sensor along the rail direction. The imaginary unit; The stress wave spectrum after compensation. This is the measured stress wave spectrum.

[0173] In a continuous medium, the wavenumber-frequency relationship is given by the following equation: ;in, Indicates frequency as The wave number corresponding to the wave, The phase velocity at this frequency varies with frequency, reflecting dispersion. Because rails have specific cross-sectional geometries (i.e., rail profile parameters, such as 60kq / m rails, 50kq / m rails, etc.), they allow the propagation of various wave types (such as flexural waves, longitudinal waves, and transverse waves). The phase velocity of each wave type is not constant, but rather a curve that varies with frequency. This phase velocity curve is obtained through numerical modeling. Based on a two-dimensional finite element model of the rail cross-section, the guided wave dispersion equation is solved to obtain the phase velocity of each mode. The curve is obtained Then, substitute the values ​​into the above formula to calculate. Used for frequency domain signal compensation:

[0174] In the frequency domain, multiply each frequency component by... That is, the wave has completed its propagation distance. Phase inversion compensation effectively eliminates waveform broadening and phase drift caused by dispersion, restoring the true structural response close to the source waveform.

[0175] Propagation feature quantization: Extract the following propagation feature parameters from the compensated waveform:

[0176] Time difference of arrival of wave crest: ;

[0177] Energy percentage of characteristic frequency band (50-80Hz): ;

[0178] Energy attenuation rate between adjacent sleepers: ;

[0179] Integrating the eigenvalues ​​of all nodes, a propagation sequence matrix of orbital stress waves is formed:

[0180] ;

[0181] in, Indicates the first Peak arrival time at each node Indicates the first Peak arrival time at each node Indicates the first Wave energy at each node Indicates the first Wave energy at each node Indicates the total number of sensor nodes. This represents the time difference of arrival of the wave peaks between adjacent virtual sensor nodes. This indicates the energy percentage of the 50–80Hz characteristic frequency band. This indicates the maximum upper frequency limit of the analysis band. This represents the propagation sequence matrix of orbital stress waves, with each row representing the propagation characteristics of a node. , Indicates the first Peak arrival time of each node Indicates the first The characteristic frequency band energy percentage of each node , Indicates the first The energy decay rate of each node relative to the previous node.

[0182] Thus, through the above embodiments, this application uses dispersion compensation to correct the waveform distortion of rail wave propagation, ensuring the physical fidelity of the stress wave observation spectrum, and accurately extracts multi-dimensional propagation characteristics such as time difference, energy ratio and attenuation rate, forming a track stress wave propagation sequence with clear physical meaning, providing a high-precision and interference-resistant quantitative data foundation for subsequent defect location.

[0183] In an exemplary embodiment of this application, the method further includes a step of generating a correlation matrix between defects and risks corresponding to the defect location results at the target defect points, for use in track operation and maintenance decisions. The step of generating the correlation matrix may specifically include:

[0184] Vehicle vibration mode set extracted based on dynamic simulation trajectory;

[0185] By performing resonance matching on the vehicle body vibration mode set and the track stress wave propagation sequence, the defect type of the target defect point corresponding to the defect location result is determined, and the defect type is either resonance or non-resonance.

[0186] By using the inverse relationship between frequency and vehicle speed pre-built, the critical vehicle speed threshold corresponding to the target defect point is obtained;

[0187] Based on the defect source coordinates, defect type, and critical speed threshold of the target defect point, a correlation matrix between defects and risks is generated, which is used for track operation and maintenance decisions.

[0188] To avoid resonance, a mathematical model was established that converts frequency differences into a safe vehicle speed threshold. This model reflects the inverse relationship between frequency and vehicle speed. By adjusting the vehicle's operating speed, the excitation frequency between the wheels and rails is made to avoid the resonance frequency of the vehicle body, thereby reducing the risk of resonance. The critical speed threshold for the target defect point is obtained through this mathematical model.

[0189] In this embodiment, the torsional vibration dominant frequency of the vehicle body vibration mode set is extracted. Stress wave dominant frequency in stress wave propagation sequence It is determined whether a resonance relationship exists, and the defect type of the target defect point is determined based on the resonance relationship, which is either resonance or non-resonance.

[0190] By using a pre-established inverse relationship between frequency and vehicle speed, the critical speed threshold corresponding to the target defect point is obtained. Specifically, the modeling and acquisition of critical speed can include: key calculations using the wheelset rotation frequency coefficient (P) and rail ripple wavelength (λg) to construct a transfer function; the smaller the frequency difference, the lower the critical speed; the influence of wheel wear (μ) and total mileage (Twear) on rotation frequency is also considered to improve the long-term adaptability of the model.

[0191] When the stress wave frequency is higher than the vehicle body resonance frequency ( > This indicates "super-resonance," at which point the vehicle should be decelerated; conversely, if... < Then, one should accelerate past the resonance zone.

[0192] Establish the frequency-vehicle speed inverse relationship of the wheel-rail contact system and calculate the critical operating speed for resonance avoidance. : ;in, This represents the critical speed threshold used to avoid resonance. This indicates the wavelength of the corrugations on the rail surface. The wheelset rotational frequency coefficient, after considering wear correction, is defined as: ; The diameter of the wheel. This refers to the wheel wear coefficient. This represents the total mileage traveled.

[0193] Note: when When the time indicates "super resonance," the speed should be reduced; conversely, the speed should be increased.

[0194] Finally, the risk assessment results for each defect are output, forming a defect-risk correlation matrix, which includes:

[0195] Defect source coordinates: latitude and longitude + chain number:

[0196] Associated risk type: Resonance / Non-resonance;

[0197] Critical vehicle speed threshold .

[0198] In addition, after obtaining the correlation matrix between defects and risks, the steps can also include outputting the maintenance coordinate set and critical speed: sorting and labeling all coordinates according to the defect risk level.

[0199] Level 1 maintenance coordinates (red label): High-risk resonance defect;

[0200] Level 2 maintenance coordinates (yellow label): Abnormal energy attenuation but no resonance characteristics;

[0201] Simultaneously output the critical speed corresponding to each defect point. The geofence module is written into the vehicle control system for automatic speed limiting and risk warning during operation.

[0202] In railway operation and maintenance, maintenance resources are limited and time windows are restricted. Especially on busy lines or mountain tunnel sections, it is necessary to accurately formulate a strategy of "where to repair first and where to repair later". Therefore, it is necessary to introduce a risk level classification mechanism to make the maintenance priority of different defects explicit, so as to avoid the problems of resource misallocation or high-risk missed detection caused by the traditional "processing according to the order of discovery".

[0203] As can be seen from the waveguide location and resonance risk assessment above, not all defects have the same impact on operational safety: Resonance defects, such as those where the torsional vibration frequency matches the structural response, can cause a large vibration amplification, which is sudden and highly damaging, and must be dealt with first; Non-resonance defects with abnormal energy attenuation, although they will not immediately cause structural resonance, are often precursors to early fatigue or track loosening, and also need to be tracked and repaired.

[0204] Therefore, this application classifies defects according to physical risk mechanisms, using technical means to support the scientific and visual nature of maintenance decisions. Traditional defect detection schemes often only output "a defect has been found" without providing specific operational response strategies. This application introduces a critical vehicle speed (Vc) as a "control variable" for operation management, and combines it with "defect coordinates" to output to a geofencing module. This enables the vehicle to automatically reduce speed when approaching a high-risk defect point and resume normal speed after leaving the risk area. This achieves an integrated closed loop of defect identification → control commands → real-time speed limiting. This allows this application to not only serve maintenance scheduling but also directly support the implementation of a dynamic operation safety control system.

[0205] Thus, through the above embodiments, this application identifies high-risk resonance defects using a resonance matching mechanism and derives the corresponding critical safe speed threshold based on a wheel-rail dynamics model. Based on this, it generates a correlation matrix that integrates defect location, risk type, and speed limit instructions. This matrix directly transforms structural health diagnostic results into actionable operational and maintenance decisions, achieving a closed loop from defect detection to risk warning and proactive control, significantly improving the accuracy and proactivity of track safety management.

[0206] Furthermore, based on defect identification, a maintenance coordinate set including latitude and longitude, chain number, risk level, and critical speed is generated. This coordinate set is then tagged and graded according to resonance risk priority, allowing direct integration with the geofencing modules of the civil engineering maintenance system and the vehicle control system. This enables a closed-loop process from trajectory data acquisition to simulation modeling, modal analysis, defect inversion, risk assessment, and control output. Compared to traditional management methods relying on manual screening and static reports, this improves the accuracy of maintenance scheduling, the timeliness of risk response, and the automation level of operational control, while also demonstrating excellent engineering adaptability and system integration capabilities.

[0207] In another exemplary embodiment, the dynamic simulation trajectory also includes the contact envelope between the wheel tread and the track section. Therefore, in the scenario described above, the step of extracting the vehicle vibration modes based on the dynamic simulation trajectory may specifically include:

[0208] Based on the contact envelope, the longitudinal excitation frequency and lateral excitation amplitude of the vehicle body are obtained.

[0209] Using a pre-constructed multibody dynamics model of the vehicle, the phase angle of the first-order torsional vibration mode and the wheel-rail adhesive oscillation entropy are obtained based on the longitudinal excitation main frequency and the lateral excitation amplitude.

[0210] Based on the longitudinal excitation main frequency, the lateral excitation amplitude, the phase angle of the first-order torsional vibration mode, and the wheel-rail adhesive oscillation entropy, the vehicle body vibration mode set is obtained.

[0211] Based on the above embodiments, an exemplary embodiment provided in this application mainly includes the steps of extracting the vehicle vibration mode set, modal decoupling modeling, and outputting the vehicle vibration mode set.

[0212] Vehicle vibration mode set extraction: Based on the wheel-rail contact envelope generated from the dynamic simulation trajectory, the dynamic excitation experienced by the vehicle body is extracted, and its response is modally decomposed by establishing a dynamic model, specifically including:

[0213] Based on the wheel-rail contact envelope in the dynamic simulation trajectory, the excitation characteristics of the vehicle body are constructed:

[0214] Vertical excitation frequency: ;in, The curvature function representing the contact point. This indicates the dominant frequency caused by changes in excitation.

[0215] Lateral excitation amplitude: ;in, This indicates the change in the lateral offset of the contact envelope over time.

[0216] Modal decoupling modeling: Establish a multibody dynamics model, set the vehicle body flexibility parameters, and use... To incentivize the input, the following response metrics are calculated:

[0217] Phase angle of first-order torsional vibration mode: ;in, Describes the shape function of the first-order torsional vibration mode. This indicates the corresponding phase angle.

[0218] Wheel-rail adhesive oscillation entropy (based on frequency domain power spectrum): ;in, This represents the normalized probability distribution of the power spectrum of the wheel-rail interaction force. This represents a frequency domain uncertainty index that reflects the wheel-rail coupling adhesion state.

[0219] Final output vehicle body vibration mode set: .

[0220] Curvature function of contact point Specifically, it includes:

[0221] Normal displacement of wheel-rail contact point : The contact point normal response (a slight vertical jump) obtained from explicit dynamic simulation;

[0222] Rail axial displacement : The displacement trajectory corresponding to the rolling direction of the wheel;

[0223] Approximate calculation of curvature: based on difference or derivative approximation. It can be calculated in the following form: Alternatively, it can be represented in the time series using a sliding window approach: ;in, The normal position of the contact point. For unit wheel displacement, This represents the corresponding time interval.

[0224] Thus, through the above embodiments, this application accurately extracts the longitudinal and lateral excitation features acting on the vehicle body based on a high-fidelity dynamic contact envelope, and inputs them into a multibody dynamics model to decouple the first-order torsional phase angle characterizing the key dynamic response of the vehicle body and the adhesive oscillation entropy reflecting the stability of the wheel-rail interface. The resulting vehicle body vibration mode set not only quantifies the vibration state of the vehicle body itself, but also profoundly reveals the dynamic interaction of the vehicle-rail coupling, providing a complete and reliable physical parameter basis for the subsequent accurate identification of vehicle body resonance risks caused by track defects.

[0225] In another exemplary embodiment, the step of determining the defect type of the target defect point by performing resonance matching on the vehicle body vibration mode set and the track stress wave propagation sequence may specifically include:

[0226] The torsional vibration dominant frequency of the target defect point in the vehicle body vibration mode set and the stress wave dominant frequency in the track stress wave propagation sequence were extracted.

[0227] If the torsional vibration frequency and the stress wave frequency satisfy the preset resonance condition, the defect type of the target defect point corresponding to the stress wave frequency is determined to be resonance, and the defect type of other target defect points is non-resonance.

[0228] In this embodiment, by comparing the torsional vibration dominant frequency of the vehicle body vibration mode set ( ) and the dominant frequency of stress waves in the orbital stress wave propagation sequence ( This involves determining whether the structures resonate. Resonance occurs when the natural frequency of a structure is close to the frequency of an external excitation, which can lead to response amplification and structural damage.

[0229] The specific judgment criteria can include two:

[0230] Frequency proximity: When the difference between the dominant frequency of torsional vibration and the dominant frequency of stress wave is less than or equal to 5Hz;

[0231] Phase alignment: The phase difference between the two is less than or equal to π / 6, indicating that they are almost synchronized in time.

[0232] If the above two conditions are met, the defect is determined to have a high-risk resonance and should be addressed as a priority.

[0233] Specifically, resonance risk assessment involves extracting the dominant torsional frequency from the vehicle body's vibration mode set. The dominant frequency of the characteristic frequency band of the stress wave propagation sequence To determine whether a resonance relationship exists, a high-risk resonance is defined as one that meets the following two conditions: ;in, The dominant frequency of the torsional vibration mode of the vehicle body. The center frequency of the orbital stress wave band. This represents the phase difference between the two.

[0234] Thus, through the above embodiments, this application dynamically matches the track stress wave propagation sequence with the vehicle body vibration mode set, extracts the frequency difference and phase difference features between the torsional vibration main frequency and the characteristic frequency band, determines whether the structural resonance condition is triggered, and derives a critical speed threshold model based on wheel-rail contact dynamics, fully considering the influencing factors such as wheel diameter, wear and rail surface ripples, to achieve adaptive avoidance of structural resonance risk by vehicle speed. Compared with the traditional experience-based speed limit method, it has stronger physical interpretability and dynamic adaptability, and can provide real-time speed limit reference for train operation control system, improving operation safety and scheduling intelligence.

[0235] Figure 3 A schematic diagram of the track defect location device provided in this application is shown below. Figure 3 As shown, the track defect locating device 30 includes:

[0236] The simulation trajectory unit 301 is used to acquire the dynamic simulation trajectory on the target track section. The dynamic simulation trajectory is generated by bidirectional coupling simulation based on the vehicle's trajectory data on the target track section and the track digital twin of the target track section. The dynamic simulation trajectory includes the time-varying energy distribution spectrum of stress wave propagation along the track.

[0237] The feature extraction unit 302 is used to extract the orbital stress wave propagation sequence of the target orbital segment based on the time-varying energy distribution spectrum. The orbital stress wave propagation sequence includes the measured energy attenuation rate of the orbital stress wave.

[0238] Inversion calculation unit 303 is used to perform inversion calculations based on the orbital stress wave propagation sequence to obtain the theoretical energy attenuation rate corresponding to candidate defect points on the target orbital segment.

[0239] The consistency verification unit 304 is used to verify the consistency between the theoretical energy decay rate and the measured energy decay rate.

[0240] The defect location unit 305 is used to take the candidate defect points that have passed the consistency check as the defect location results of the target track section.

[0241] In one possible implementation, the inversion calculation unit 303 is further used to obtain the reference characteristic frequency of the track stress wave corresponding to two adjacent reference points on the target track segment based on the track stress wave propagation sequence; the candidate defect point is located on the track between the two adjacent reference points, and the reference points are virtual sensor nodes deployed on the target track segment; the wave value at the reference characteristic frequency is obtained based on the relationship function between wave number and frequency; the relationship function corresponds to the rail type of the target track segment; the actual phase velocity corresponding to the reference characteristic frequency is calculated based on the wave value and the reference characteristic frequency; the wave peak arrival time difference between the reference characteristic frequencies corresponding to the two adjacent reference points is obtained based on the track stress wave propagation sequence; the defect source distance of the candidate defect point is calculated based on the actual phase velocity and the wave peak arrival time difference through the difference frequency propagation positioning equation, and the defect source distance is the distance between the candidate defect point and the reference point; the theoretical energy attenuation rate corresponding to the candidate defect point is calculated based on the defect source distance and the energy attenuation factor of the target track segment.

[0242] In one possible implementation, the consistency verification unit 304 is also used to extract the measured energy attenuation rate between two reference points corresponding to the candidate defect point based on the orbital stress wave propagation sequence, where the reference points are virtual sensor nodes on the target orbital segment adjacent to the candidate defect point; and to obtain the attenuation rate error between the measured energy attenuation rate and the theoretical energy attenuation rate.

[0243] The consistency verification result is obtained based on the absolute value of the attenuation rate error and the error threshold.

[0244] In one possible implementation, the simulation trajectory unit 301 is further configured to: obtain the vehicle's load on the target track segment based on the position and attitude information included in the trajectory data; retrieve the track digital twin corresponding to the target track segment from the database; apply the load to the track digital twin to obtain the elastic deformation field and initial excitation intensity of the stress wave on the track surface under the load; obtain the deformed track surface geometry based on the elastic deformation field and the three-dimensional point cloud model of the rail included in the track digital twin; obtain the contact envelope between the vehicle's wheel tread and the track based on the deformed track surface geometry using a contact geometry inversion algorithm; and generate the dynamic simulation trajectory of the track based on the contact envelope and the initial excitation intensity of the stress wave.

[0245] In one possible implementation, the feature extraction unit 302 is further configured to acquire the stress wave observation spectrum based on the time-varying energy distribution spectrum using a virtual sensor array, the virtual sensor array being axially deployed on the target track section; perform dispersion compensation on the stress wave observation spectrum based on the relationship function between the wave number and frequency corresponding to the rail type of the track, to obtain the compensated stress wave spectrum; extract features from the compensated stress wave spectrum to obtain propagation feature parameters; and form the track stress wave propagation sequence based on the propagation feature parameters.

[0246] In one possible implementation, the track defect location device further includes an association unit for extracting a set of vehicle vibration modes based on a dynamic simulation trajectory; determining the defect type of the target defect point by performing resonance matching between the set of vehicle vibration modes and the track stress wave propagation sequence, wherein the defect type is resonant or non-resonant; obtaining the critical vehicle speed threshold corresponding to the target defect point by using a pre-built inverse relationship between frequency and vehicle speed; and generating an association matrix between the defect and risk based on the defect source coordinates, defect type, and critical vehicle speed threshold of the target defect point, wherein the association matrix is ​​used for track operation and maintenance decisions.

[0247] In one possible implementation, the dynamic simulation trajectory also includes the contact envelope between the wheel tread and the track section; the associated unit is also used to obtain the longitudinal excitation dominant frequency and lateral excitation amplitude of the vehicle body based on the contact envelope; through the pre-constructed multibody dynamics model of the vehicle, the phase angle of the first-order torsional vibration mode and the wheel-rail adhesive oscillation entropy are obtained based on the longitudinal excitation dominant frequency and the lateral excitation amplitude; based on the longitudinal excitation dominant frequency, the lateral excitation amplitude, the phase angle of the first-order torsional vibration mode and the wheel-rail adhesive oscillation entropy, the vehicle body vibration mode set is obtained.

[0248] In one possible implementation, the association unit is also used to extract the torsional vibration dominant frequency of the target defect point in the vehicle body vibration mode set, and the stress wave dominant frequency in the track stress wave propagation sequence; if the torsional vibration dominant frequency and the stress wave dominant frequency meet the preset resonance condition, the defect type of the target defect point corresponding to the stress wave dominant frequency is determined to be resonance, and the defect type of other target defect points is non-resonance.

[0249] The track defect location device provided in this embodiment can execute the method provided in the above method embodiment. Its implementation principle and technical effect are similar, and will not be described in detail here.

[0250] Figure 4 A schematic diagram of the structure of the electronic device provided in this application. Figure 4 As shown, the electronic device 40 provided in this embodiment includes at least one processor 401 and a memory 402. Optionally, the device 40 further includes a communication component 403. The processor 401, memory 402, and communication component 403 are connected via a bus 404.

[0251] In a specific implementation, at least one processor 401 executes computer execution instructions stored in memory 402, causing at least one processor 401 to perform the above-described method.

[0252] The specific implementation process of processor 401 can be found in the above method embodiments, and its implementation principle and technical effect are similar. It will not be repeated here.

[0253] In the above embodiments, it should be understood that the processor can be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), etc. The general-purpose processor can be a microprocessor or any conventional processor. The steps of the method disclosed in this invention can be directly implemented by a hardware processor, or implemented by a combination of hardware and software modules within the processor.

[0254] The memory may include random access memory (RAM) and may also include non-volatile memory (NVM), such as at least one disk storage device.

[0255] The bus can be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus, or an Extended Industry Standard Architecture (EISA) bus, etc. Buses can be categorized as address buses, data buses, control buses, etc. For ease of illustration, the buses shown in the accompanying drawings are not limited to a single bus or a single type of bus.

[0256] This application also provides a computer program product, including a computer program that, when executed by a processor, implements the above-described method.

[0257] This application also provides a computer-readable storage medium storing computer-executable instructions, which, when executed by a processor, implement the above-described method.

[0258] The aforementioned readable storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk. The readable storage medium can be any available medium accessible to a general-purpose or special-purpose computer.

[0259] An exemplary readable storage medium is coupled to a processor, enabling the processor to read information from and write information to the readable storage medium. Of course, the readable storage medium can also be a component of the processor. The processor and the readable storage medium can reside in an Application Specific Integrated Circuit (ASIC). Alternatively, the processor and the readable storage medium can exist as discrete components in the device.

[0260] The division of units is merely a logical functional division; in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be indirect coupling or communication connection through some interfaces, devices, or units, and may be electrical, mechanical, or other forms.

[0261] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0262] In addition, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.

[0263] If a function is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this invention, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of the various embodiments of this invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0264] Those skilled in the art will understand that all or part of the steps of the above-described method embodiments can be implemented by hardware related to program instructions. The aforementioned program can be stored in a computer-readable storage medium. When executed, the program performs the steps of the above-described method embodiments; and the aforementioned storage medium includes various media capable of storing program code, such as ROM, RAM, magnetic disks, or optical disks.

[0265] Finally, it should be noted that other embodiments of the invention will readily occur to those skilled in the art upon consideration of the specification and practice of the invention disclosed herein. This invention is intended to cover any variations, uses, or adaptations of the invention that follow the general principles of the invention and include common knowledge or customary techniques in the art not disclosed herein, and is not limited to the precise structures described above and shown in the accompanying drawings, and various modifications and changes can be made without departing from its scope. The scope of the invention is limited only by the appended claims.

Claims

1. A method for locating track defects, characterized in that, include: The dynamic simulation trajectory on the target track section is obtained. The dynamic simulation trajectory is generated by bidirectional coupling simulation based on the vehicle's trajectory data on the target track section and the track digital twin of the target track section. The dynamic simulation trajectory includes the time-varying energy distribution spectrum of stress wave propagation along the track. Based on the time-varying energy distribution spectrum, the orbital stress wave propagation sequence of the target orbital segment is extracted, and the orbital stress wave propagation sequence includes the measured energy attenuation rate of the orbital stress wave. Inversion calculations are performed based on the orbital stress wave propagation sequence to obtain the theoretical energy attenuation rate corresponding to the candidate defect point on the target orbital segment. The consistency between the theoretical energy decay rate and the measured energy decay rate is verified. Candidate defect points that pass the consistency check are used as the defect location results for the target track segment.

2. The track defect location method according to claim 1, characterized in that, The inversion calculation based on the orbital stress wave propagation sequence to obtain the theoretical energy attenuation rate corresponding to the candidate defect point on the target orbital segment includes: Based on the orbital stress wave propagation sequence, the reference characteristic frequencies of the orbital stress waves corresponding to two adjacent reference points on the target orbital section are obtained; the candidate defect points are located on the track between two adjacent reference points, and the reference points are virtual sensor nodes deployed on the target orbital section. Based on the relationship function between wavenumber and frequency, the wave value at the reference characteristic frequency is obtained; the relationship function corresponds to the rail type of the target track section. Based on the wave value and the reference characteristic frequency, the actual phase velocity corresponding to the reference characteristic frequency is calculated; Based on the orbital stress wave propagation sequence, the time difference of wave peak arrival between the reference characteristic frequencies corresponding to two adjacent reference points is obtained. The defect source distance of the candidate defect point is calculated based on the actual phase velocity and the wave peak arrival time difference using the difference frequency propagation positioning equation. The defect source distance is the distance between the candidate defect point and the reference point. Based on the distance to the defect source and the energy attenuation factor of the target orbital segment, the theoretical energy attenuation rate corresponding to the candidate defect point is calculated.

3. The track defect location method according to claim 1, characterized in that, The consistency verification between the theoretical energy decay rate and the measured energy decay rate includes: Based on the orbital stress wave propagation sequence, the measured energy attenuation rate between two reference points corresponding to the candidate defect point is extracted. The reference point is a virtual sensor node on the target orbital segment adjacent to the candidate defect point. Obtain the decay rate error between the measured energy decay rate and the theoretical energy decay rate; The consistency verification result is obtained based on the absolute value of the attenuation rate error and the error threshold.

4. The track defect location method according to any one of claims 1 to 3, characterized in that, The acquisition of the dynamic simulation trajectory on the target track segment includes: Based on the position and attitude information included in the trajectory data, the load of the vehicle on the target track section is obtained; Retrieve the digital twin of the target track segment from the database; The load is applied to the digital twin of the track to obtain the elastic deformation field of the track surface and the initial excitation intensity of the stress wave under the load; Based on the elastic deformation field and the three-dimensional point cloud model of the rail included in the digital twin of the track, the geometric state of the rail surface after deformation is obtained. The contact envelope between the vehicle's wheel tread and the track is obtained based on the deformed track surface geometry using a contact geometry inversion algorithm. Based on the contact envelope and the initial excitation intensity of the stress wave, a dynamic simulation trajectory of the target orbital segment is generated.

5. The method for locating track defects according to any one of claims 1 to 3, characterized in that, The extraction of the orbital stress wave propagation sequence of the target orbital segment based on the time-varying energy distribution spectrum includes: The stress wave observation spectrum is obtained by a virtual sensor array based on the time-varying energy distribution spectrum, and the virtual sensor array is axially arranged on the target orbital segment. Based on the relationship function between the wave number and frequency corresponding to the rail type of the track, the stress wave observation spectrum is subjected to dispersion compensation to obtain the compensated stress wave spectrum. The propagation characteristic parameters are obtained by extracting features from the compensated stress wave spectrum. The orbital stress wave propagation sequence of the target orbital segment is formed based on the propagation characteristic parameters.

6. The method for locating track defects according to any one of claims 1 to 3, characterized in that, Also includes: The vehicle vibration mode set is extracted based on the dynamic simulation trajectory; By performing resonance matching on the vehicle body vibration mode set and the track stress wave propagation sequence, the defect type of the target defect point corresponding to the defect location result is determined, and the defect type is either resonance or non-resonance. By using a pre-established inverse relationship between frequency and vehicle speed, the critical vehicle speed threshold corresponding to the target defect point is obtained; Based on the defect source coordinates, defect type, and critical speed threshold of the target defect point, a correlation matrix between defects and risks is generated, and the correlation matrix is ​​used for track operation and maintenance decisions.

7. The track defect location method according to claim 6, characterized in that, The dynamic simulation trajectory also includes the contact envelope between the wheel tread and the track section; the extraction of the vehicle vibration mode set based on the dynamic simulation trajectory includes: Based on the contact envelope, the longitudinal excitation frequency and lateral excitation amplitude of the vehicle body are obtained. Based on the pre-constructed multibody dynamics model of the vehicle, the phase angle of the first-order torsional vibration mode and the wheel-rail adhesive oscillation entropy are obtained using the longitudinal excitation main frequency and the lateral excitation amplitude. Based on the longitudinal excitation main frequency, the lateral excitation amplitude, the phase angle of the first-order torsional vibration mode, and the wheel-rail adhesion oscillation entropy, the vehicle body vibration mode set is obtained.

8. The track defect location method according to claim 7, characterized in that, The method of determining the defect type of the target defect point by performing resonance matching on the vehicle body vibration mode set and the track stress wave propagation sequence includes: The torsional vibration dominant frequency of the target defect point in the vehicle body vibration mode set and the stress wave dominant frequency in the track stress wave propagation sequence are extracted. If the torsional vibration frequency and the stress wave frequency satisfy the preset resonance condition, the defect type of the target defect point corresponding to the stress wave frequency is determined to be resonance, and the defect type of other target defect points is non-resonance.

9. A track defect locating device, characterized in that, include: The simulation trajectory unit is used to acquire the dynamic simulation trajectory on the target track section. The dynamic simulation trajectory is generated by bidirectional coupling simulation based on the vehicle's trajectory data on the target track section and the track digital twin of the target track section. The dynamic simulation trajectory includes the time-varying energy distribution spectrum of stress wave propagation along the track. The feature extraction unit is used to extract the orbital stress wave propagation sequence of the target orbital segment based on the time-varying energy distribution spectrum, wherein the orbital stress wave propagation sequence includes the measured energy attenuation rate of the orbital stress wave; The inversion calculation unit is used to perform inversion calculations based on the orbital stress wave propagation sequence to obtain the theoretical energy attenuation rate corresponding to the candidate defect point on the target orbital segment. A consistency verification unit is used to verify the consistency between the theoretical energy decay rate and the measured energy decay rate. The defect location unit is used to take the candidate defect points that pass the consistency check as the defect location results of the target track segment.

10. An electronic device, characterized in that, include: One or more processors; A storage device for storing one or more programs, which, when executed by one or more processors, cause the electronic device to implement the track defect location method as described in any one of claims 1 to 8.

Images