A method, apparatus and system for synergistically triggering compensation for monitoring sleepers

Abstract

The application discloses a kind of cooperative triggering compensation type sleeper monitoring method, device and system.The monitoring method includes real-time monitoring strain signal and vibration signal under sleeper, and obtains the first temperature of monitoring area;Obtain the temperature and humidity of external environment, based on the temperature and humidity of external environment and the first temperature respectively to the strain signal, the vibration signal is compensated, and strain compensation signal and vibration compensation signal are obtained;Based on the strain compensation signal and vibration compensation signal, extract multiple characteristics, and utilize the feature to calculate event confidence;According to the event confidence, the construction of track section event chain is triggered, and according to the track section event chain, event type is judged.The present application can improve the long-term monitoring accuracy of strain under rail, vibration and environmental data, enhance the capture ability of impact event and supporting state anomaly, and realize the cooperative identification of abnormal position and abnormal type of track section.

Description

Technical Field

[0001] This invention relates to the field of monitoring technology for rail transit infrastructure, and more specifically to a collaborative triggering compensation sleeper monitoring method, device, and system. Background Technology

[0002] As a crucial component of the track structure responsible for bearing and transmitting train loads, the sleeper's bearing area, the underlying pad, and the support condition of the sleeper bottom directly affect the track's geometric stability and the safety of the line's service life. With the long-term, high-density operation of high-speed railways, heavy-haul railways, and urban rail transit, sleepers and their adjacent structures are prone to problems such as localized abnormal stress, changes in support stiffness, increased impact vibration, moisture intrusion, and material aging. To monitor the service condition of the underlying structure, current technologies employ strain sensors, acceleration sensors, temperature and humidity sensors, and wireless data acquisition nodes for online monitoring of sleepers or the area beneath them.

[0003] However, existing monitoring solutions typically involve installing multiple sensors on the surface of the sleeper, near the rail subbase, or inside a separate housing, and then uploading the data via acquisition circuitry and a communication module. While this approach can obtain certain mechanical and environmental parameters, it still has the following shortcomings:

[0004] 1. The sensor lacks a stable and clear mechanical coupling structure with the force path under the rail, and the monitoring results are easily affected by the installation position, bonding status, loose housing, and local contact conditions.

[0005] 2. Temperature and humidity data are mostly used as supplementary records and fail to effectively compensate for strain zero-point drift, sensitivity changes, and vibration response drift, resulting in a long-term decrease in monitoring accuracy;

[0006] 3. Event triggering often relies on a single vibration amplitude or a fixed threshold, making it difficult to distinguish between train load, accidental impact, environmental disturbance, and node anomalies, which can easily lead to false triggering or missed triggering.

[0007] 4. Multiple monitoring nodes typically only aggregate data, lacking spatial collaborative discrimination based on the time delay, amplitude attenuation, and response consistency of adjacent nodes, making it difficult to accurately identify abnormal locations and changes in the track support status.

[0008] Therefore, there is an urgent need for a smart sleeper structure or monitoring method that can take into account track-ground mechanical coupling, environmental isolation compensation, multi-source event identification, and multi-node collaborative diagnosis. Summary of the Invention

[0009] In view of the above problems, the present invention proposes a collaborative triggering compensation sleeper monitoring method, device and system, aiming to improve the long-term monitoring accuracy of track strain, vibration and environmental data, enhance the ability to capture impact events and support state anomalies, and realize collaborative identification of anomaly locations and types in track sections.

[0010] To achieve the above objectives, the present invention adopts the following technical solution:

[0011] In a first aspect, embodiments of the present invention provide a collaboratively triggered compensation-type sleeper monitoring method, the steps of which include:

[0012] Real-time monitoring of strain and vibration signals under the sleepers, and acquisition of the first temperature of the monitoring area;

[0013] The ambient temperature and humidity are acquired, and the strain signal and the vibration signal are compensated based on the ambient temperature and humidity and the first temperature to obtain the strain compensation signal and the vibration compensation signal.

[0014] Multiple features are extracted based on the strain compensation signal and vibration compensation signal, and the event confidence is calculated using the features;

[0015] The event confidence level is used to trigger the construction of the track segment event chain, and the event type is determined based on the track segment event chain.

[0016] Preferably, compensating for the strain signal based on the external ambient temperature and humidity and the first temperature includes:

[0017]

[0018] In the formula, Indicates strain compensation signal, This represents the strain signal monitored in real time. Indicates the strain signal of the reference channel. Indicates the reference channel compensation coefficient. Indicates the equivalent temperature of the monitored area. Indicates the reference temperature of the monitoring area. This represents the temperature difference between the initial temperature and the ambient temperature. Indicates the humidity of the external environment. This represents the strain compensation coefficient.

[0019] Preferably, compensating for the vibration signal based on the external ambient temperature and humidity and the first temperature includes:

[0020]

[0021] In the formula, Indicates vibration compensation signal, This indicates the vibration signal being monitored in real time. The equivalent temperature of the monitoring area Indicates the reference temperature of the monitoring area. Indicates the humidity of the external environment. This represents the vibration compensation coefficient.

[0022] Preferably, the strain compensation coefficient and vibration compensation coefficient are updated online using recursive least squares, moving average, or exponential smoothing methods; and the update of the compensation coefficient is frozen under event conditions.

[0023] Preferably, the equivalent temperature of the monitoring area is obtained in the following manner:

[0024]

[0025] In the formula, Represents the smoothing coefficient. Indicates the first temperature. Indicates the ambient temperature;

[0026] or

[0027]

[0028] In the formula, t represents the current sampling time. Indicates based on the first temperature relative to external ambient temperature The temperature lag time determined by the rate of change or historical correlation, This indicates the effective temperature sampling time after considering thermal inertia hysteresis.

[0029] Preferably, the multiple features are extracted based on a sliding time window, including strain-vibration time correlation, and

[0030] Strain peak value and / or strain rate of change extracted from strain compensation signal;

[0031] The vibration peak value, root mean square value, rate of change of acceleration, impact kurtosis and / or frequency band energy are extracted from the vibration compensation signal.

[0032] Preferably, the normalized cross-correlation curves of the strain compensation sequence and the vibration compensation sequence are calculated within a preset time delay range, and the maximum absolute correlation coefficient of the cross-correlation curve is taken as the strain-vibration time correlation.

[0033] Preferably, calculating the event confidence using the features includes:

[0034] Determine the mean and standard deviation of the background baseline corresponding to each feature;

[0035] The current feature is normalized based on the mean and standard deviation.

[0036] The normalized features are weighted and summed.

[0037] Preferably, the construction of the track segment event chain is triggered based on the event confidence level, and the event type is determined based on the track segment event chain, including:

[0038] When the confidence level of the event is higher than a preset threshold, a mutual trigger command is sent to a preset number of neighboring monitoring points.

[0039] Receive the feature summaries returned by the neighboring monitoring points in response to the mutual triggering command, and sort them according to the response arrival time and node location to form a track segment event chain;

[0040] The event type is determined based on the relationship between the features in the feature summary of each event chain and the change of the node position.

[0041] As a preferred approach, if the event response is concentrated on a single sleeper or a few adjacent sleepers, it is judged as a local impact or local support anomaly; if the event response propagates regularly along multiple sleepers, the direction of impact propagation or train passing characteristics can be determined by combining the trigger time sequence.

[0042] The feature summary includes node location, timestamp, event confidence, vibration amplitude, compensation strain value, and correlation C.

[0043] Preferably, when the event confidence level is lower than a preset threshold, a first sampling strategy is used for monitoring; otherwise, a second sampling strategy is used for monitoring, wherein the energy consumption of the first sampling strategy is lower than that of the second sampling strategy.

[0044] Preferably, a switching threshold is set, which is lower than a preset threshold. When the event confidence is lower than the switching threshold for a preset duration, the system reverts to the first sampling strategy to form a hysteresis interval and reduce repeated switching near the threshold.

[0045] Secondly, embodiments of the present invention provide a collaborative triggering compensation sleeper monitoring device, the device comprising: an elastic measuring unit and a force coupling seat;

[0046] The elastic measurement unit is fixed above the force coupling seat and includes an environmental reference cavity and a sealed acquisition cavity inside; a strain acquisition module and a vibration acquisition module are provided on the top.

[0047] The strain acquisition module is used to acquire the strain response transmitted to the elastic measurement unit via the force coupling seat;

[0048] The vibration acquisition module is used to acquire the vibration response transmitted to the elastic measuring unit via the force coupling seat;

[0049] The environmental acquisition chamber includes a first environmental unit and a second environmental unit; the first environmental unit is used to acquire a first temperature of the monitoring area, and the second environmental unit is used to acquire the external environmental temperature and humidity.

[0050] Preferably, the environmental reference cavity is connected to the outside air through a labyrinthine waterproof and breathable channel, and is isolated from the force coupling seat by a moisture-proof and vibration-damping component.

[0051] The sealed acquisition cavity is equipped with a main control module, which is used to compensate the strain signal and the vibration signal based on the external ambient temperature and humidity and the first temperature, respectively, to obtain strain compensation signal and vibration compensation signal; extract various features based on the strain compensation signal and vibration compensation signal, and use the features to calculate the event confidence level; and trigger the construction of the track segment event chain according to the event confidence level, and determine the event type according to the track segment event chain.

[0052] Thirdly, embodiments of the present invention provide a collaborative triggering compensation sleeper monitoring system, which includes multiple collaborative triggering compensation sleeper monitoring devices as described above, and a host computer;

[0053] The host computer is used to receive feature summaries uploaded by each collaboratively triggered compensating sleeper monitoring device to construct a track section event chain, and to determine the event type based on the track section event chain.

[0054] This invention provides a collaborative triggering compensation sleeper monitoring method, device, and system, which, compared with the prior art, offer at least the following advantages:

[0055] 1. This invention does not rely on a single fixed threshold trigger. Instead, it calculates event confidence based on strain characteristics, vibration characteristics, environmental conditions, and strain-vibration time correlation. Combined with dynamic baseline, hysteresis criterion, and pre-trigger cache, it can improve the completeness and reliability of abnormal event capture.

[0056] 2. By using a hierarchical control method of low-power inspection sampling, verification sampling, and high-frequency event sampling, effective data before, during, and after critical events are retained while reducing long-term operating energy consumption.

[0057] 3. By combining dual-environment unit, temperature hysteresis correction model, non-event window self-updating and event state freezing mechanism, the train load or impact response is not mistakenly learned as environmental drift, thus improving the stability of long-term strain and vibration monitoring.

[0058] 4. By using the first and second environmental units to obtain the internal temperature and external reference temperature and humidity respectively, and combining temperature hysteresis correction and temperature and humidity coupling compensation, the impact of temperature drift, humidity intrusion, encapsulation thermal inertia and installation coupling state changes on the monitoring results can be reduced.

[0059] 5. By isolating the mechanical transmission path from the environmental reference path, the environmental reference cavity does not participate in the main force transmission. At the same time, the temperature difference between the first environmental unit and the second environmental unit is used to reflect the thermal hysteresis state of the shell, providing a data basis for temperature hysteresis correction and temperature and humidity coupling compensation from the structural level, which is different from ordinary dual-cavity waterproof shells.

[0060] 6. By analyzing the proximity-triggered interactions between multiple smart sleeper structures and the event chain of track sections, the system is no longer limited to data aggregation, but can further identify impact locations, changes in support status, and local stress anomalies, thereby improving the diagnostic capabilities of online monitoring of track infrastructure. Attached Figure Description

[0061] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort.

[0062] Figure 1 This is a schematic diagram of the structure of the collaborative triggering compensation sleeper monitoring device provided in this embodiment of the invention;

[0063] Figure 2 This is a schematic diagram of the elastic measurement unit in the collaborative triggering compensation sleeper monitoring device provided in this embodiment of the invention;

[0064] Figure 3 This is a flowchart of the collaborative triggering compensation sleeper monitoring method provided in this embodiment of the invention. Detailed Implementation

[0065] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0066] To address the problems in existing track-under-rail monitoring schemes, such as unstable coupling between sensors and the track-under-rail force path, the inability to effectively compensate for measurement drift due to environmental temperature and humidity being only recorded as supplementary data, the susceptibility to misjudgment or omission of abnormal events using a single threshold triggering method, and the lack of a spatial collaborative diagnostic mechanism among multiple monitoring nodes, this invention discloses a collaborative triggering compensation-type sleeper monitoring method, device, and system. The aim is to improve the long-term monitoring accuracy of track-under-rail strain, vibration, and environmental data, enhance the ability to capture impact events and abnormal support conditions, and achieve collaborative identification of abnormal locations and types in track sections. Specific embodiments are described below.

[0067] Example 1

[0068] The present invention first provides a collaborative triggering compensation sleeper monitoring device, which is installed near the rail bearing area of ​​a concrete sleeper, near the rail pad layer, or at a reserved installation location at the bottom of the sleeper.

[0069] The device includes an elastic measuring unit 2 and a force coupling seat 1; as shown in the figure. Figure 1 and Figure 2 As shown, the force coupling seat 1 is mechanically connected to the sleeper. For example, the force coupling seat is made of metal, fiber-reinforced composite material or high-strength engineering material, and its bottom surface is in contact with the adjacent position of the sleeper bearing area, the adjacent area of ​​the track pad layer or the reserved installation part at the bottom of the sleeper.

[0070] The elastic measuring unit 2 is fixed above the force coupling seat 1 and is externally equipped with a metal shell. Optionally, the elastic measuring unit 2 may be a thinned beam, cantilever beam, ring measuring beam, or locally weakened measuring area disposed on the force coupling seat.

[0071] Furthermore, the elasticity measurement unit 2 is equipped with an environmental reference cavity 22 and a sealed acquisition cavity 21, with the two cavities isolated from each other;

[0072] In some implementations, the environmental acquisition cavity 22 includes a first environmental unit and a second environmental unit. Through the cooperation of the first and second environmental units, the thermal hysteresis of the device body and changes in the external environment can be distinguished, thereby providing basic data for subsequent compensation processing. In this embodiment, the first environmental unit is located in the vicinity of the elastic measurement unit and includes a temperature sensor or a thermistor to acquire the first temperature of the monitoring area. The second environmental unit includes a temperature sensor, a humidity sensor, or an integrated temperature and humidity sensor, located within the environmental acquisition cavity, to acquire the external environmental temperature and humidity. Preferably, the environmental reference cavity is connected to the outside air through a labyrinthine waterproof and breathable channel 25. The labyrinthine waterproof and breathable channel may include at least two folding channel sections, a waterproof and breathable membrane, and a vent facing the non-pressure-bearing side. Simultaneously, it is isolated from the force coupling seat by a moisture-proof and vibration-damping component, so that the environmental reference cavity does not serve as the main force transmission path. More preferably, the connection point between the elastic measurement unit 2 and the force coupling seat 1 is located below the sealed acquisition cavity.

[0073] In some implementation schemes, the sealed acquisition cavity 21 houses a main control module 27, a signal conditioning circuit, a communication interface circuit, and a power supply module 26. The main control module is used to compensate for and coordinate the triggering of monitoring data; the specific data processing procedure is described in Embodiment 3.

[0074] Furthermore, a strain acquisition module 23 and a vibration acquisition module 24 are provided above the elastic measurement unit 2;

[0075] The strain acquisition module 23 is located in the strain concentration area of ​​the elastic measurement unit 1 and is used to acquire the strain response transmitted to the elastic measurement unit 1 through the force coupling seat. This structure is different from the method of randomly pasting strain gauges on the inner wall of the housing. It defines a stable force transmission path through the force coupling seat and the elastic measurement unit.

[0076] In some implementations, the strain acquisition module 23 includes a strain-sensitive branch and a temperature compensation branch. The strain-sensitive branch is located in the strain concentration area of ​​the elastic measurement unit to simultaneously sense the response caused by the under-rail load and environmental factors. It includes a strain-sensitive element, a bridge measurement unit, an amplification unit, and an analog-to-digital conversion unit. The strain-sensitive element can be a resistance strain gauge, a fiber optic grating sensing element, a piezoresistive sensing element, or an equivalent mechanical response sensing element. The temperature compensation branch is located in a reference area mechanically isolated from the under-rail load and uses the same or similar sensing material, adhesive layer, encapsulation layer, and bridge connection method as the strain-sensitive branch. It primarily reflects non-load drift caused by temperature changes, humidity effects, and changes in encapsulation materials.

[0077] The vibration acquisition module 24 is used to acquire the vibration response transmitted to the elastic measuring unit via the force coupling seat. A triaxial accelerometer unit can be used, which is mounted on a vibration mounting base rigidly connected to the elastic measuring unit to acquire vertical, lateral, and longitudinal vibration responses. The vibration acquisition module can also use a velocity sensor unit, impact sensor unit, or inertial measurement unit capable of characterizing vibration properties.

[0078] This application differs from ordinary dual-cavity waterproof shells by isolating the mechanical transmission path from the environmental reference path, so that the environmental reference cavity does not participate in the main force transmission. At the same time, it uses the temperature difference between the first environmental unit and the second environmental unit to reflect the thermal hysteresis state of the shell, providing a data basis for temperature hysteresis correction and temperature and humidity coupling compensation from a structural level.

[0079] Example 2

[0080] This embodiment further provides a collaborative triggering compensation sleeper monitoring system, including multiple of the above-mentioned collaborative triggering compensation sleeper monitoring devices, and at least one aggregation node or host computer.

[0081] Multiple collaboratively triggered compensating sleeper monitoring devices are deployed along the track line and upload monitoring data to a convergence node or host computer via a communication link. When any collaboratively triggered compensating sleeper monitoring device enters an event state, it can trigger neighboring nodes to enter a verification sampling state. The convergence node or host computer forms a track section event chain based on the event confidence, response arrival time, spatial attenuation relationship, and strain response distribution of multiple nodes to identify impact locations, abnormal support sections, or localized abnormal stress on sleepers. The specific process is described in Example 3.

[0082] It should be noted that when a certain collaborative triggering compensation sleeper monitoring device is used as a convergence node, the main control module constructs the track section event chain and determines the force anomaly; otherwise, the host computer executes the command and control.

[0083] Example 3

[0084] This embodiment specifically provides a collaborative triggering compensation sleeper monitoring method, such as... Figure 3 As shown, the steps include:

[0085] S1. Real-time monitoring of strain and vibration signals under the sleepers, and acquisition of the first temperature of the monitoring area;

[0086] S2. Obtain the ambient temperature and humidity, and compensate the strain signal and the vibration signal based on the ambient temperature and humidity and the first temperature to obtain the strain compensation signal and the vibration compensation signal.

[0087] Because the strain-sensitive element, force coupling base, and electronic circuitry have thermal inertia, simply using external temperature for compensation can easily lead to phase lag. Therefore, this embodiment uses a first temperature and the ambient temperature and humidity to jointly determine the equivalent temperature of the monitoring area. In some implementation schemes, the following methods are used to obtain the information:

[0088]

[0089] In the formula, Represents the smoothing coefficient. Indicates the first temperature. Indicates the ambient temperature;

[0090] or

[0091]

[0092] In the formula, t represents the current sampling time. Indicates based on the first temperature relative to external ambient temperature The temperature lag time determined by the rate of change or historical correlation, This indicates the effective temperature sampling time after considering thermal inertia hysteresis.

[0093] Furthermore, the strain signal is compensated using the following compensation model:

[0094]

[0095] In the formula, Indicates strain compensation signal, This represents the strain signal monitored in real time. Indicates the strain signal of the reference channel. Indicates the reference channel compensation coefficient. Indicates the equivalent temperature of the monitored area. Indicates the reference temperature of the monitoring area. This represents the temperature difference between the initial temperature and the ambient temperature. Indicates the humidity of the external environment. This represents the strain compensation coefficient.

[0096] Furthermore, the vibration signal is compensated using the following model to reduce the impact of temperature-induced zero-point shift, sensitivity variations, and humidity-induced changes in installation coupling state on event judgment:

[0097]

[0098] In the formula, Indicates vibration compensation signal, This indicates the vibration signal being monitored in real time. The equivalent temperature of the monitoring area Indicates the reference temperature of the monitoring area. Indicates the humidity of the external environment. This represents the vibration compensation coefficient.

[0099] The compensation coefficients can be obtained through factory calibration or updated online during non-event windows. During on-site updates, the main control module selects time windows where the vibration peak value, root mean square vibration value, and strain rate of change are all below the background stability threshold as non-event windows, and updates the compensation coefficients using recursive least squares, moving average, or exponential smoothing methods. The compensation coefficients are frozen during event states to prevent the actual train load response from being mistakenly learned as environmental drift.

[0100] S3. Extract various features based on the strain compensation signal and vibration compensation signal, and use the features to calculate the event confidence level;

[0101] In this embodiment, multiple features are extracted based on a sliding time window, including strain-vibration time correlation, strain peak value and / or strain rate of change extracted based on strain compensation signal; vibration peak value, vibration root mean square, acceleration rate of change, impact kurtosis and / or frequency band energy extracted based on vibration compensation signal.

[0102] When calculating the confidence level, at least three of the above features should be extracted.

[0103] In some implementation schemes, the strain-vibration time correlation C can be calculated from the strain compensation sequence and the vibration compensation sequence within a preset sliding time window. Specifically, the normalized cross-correlation curves of the two are calculated within a preset time delay range, and the maximum absolute correlation coefficient of the cross-correlation curve is taken as C, with C ranging from 0 to 1. The time delay corresponding to the maximum correlation coefficient can be used as a consistency criterion for determining whether the strain response and vibration response arrive synchronously. When C is greater than a preset correlation threshold and the corresponding time delay falls within the allowable time delay range, the strain response and vibration response are deemed to meet the consistency condition.

[0104] Furthermore, within a sliding time window, the maximum amplitude value in the vibration compensation sequence is taken as the vibration peak value Ap, the square root of the vibration square mean value within the time window is taken as the vibration root mean square Ar, the ratio of the acceleration difference between adjacent sampling points to the sampling interval or its maximum value is taken as the acceleration change rate J, the ratio of the fourth central moment of the vibration sequence to the square of the second central moment is taken as the impact kurtosis K, and the vibration sequence is subjected to frequency band filtering or spectral transformation to obtain the preset frequency band energy Eb.

[0105] For the strain compensation sequence, the maximum absolute offset is taken as the strain peak value Ep, and the ratio of the difference in compensation strain between adjacent sampling points to the sampling interval or its maximum value is taken as the strain change rate Es.

[0106] Therefore, the temperature-compensated strain and vibration data are converted into multi-source features for event confidence calculation.

[0107] In some implementations, when calculating the confidence level of an event, the mean and standard deviation of the background baseline corresponding to each feature are first determined; the current feature is normalized based on the mean and standard deviation; and then the normalized features are weighted and summed.

[0108] That is, for any feature After normalization, it becomes: ,in and The mean and standard deviation of the corresponding background baseline;

[0109] The weighted summation is expressed as: , , As weight.

[0110] Alternatively, decision trees, lightweight classification models, or rule-based models can be used to calculate the event confidence level. The event state is entered only when the event confidence level S reaches the entry threshold, or simultaneously reaches the threshold and the strain-vibration time correlation meets the consistency condition.

[0111] S4. After entering the event state, the construction of the track segment event chain is triggered according to the event confidence level, and the event type is determined according to the track segment event chain.

[0112] In this embodiment, upon entering the event state, the sampling strategy is first switched from a first sampling strategy to a second sampling strategy, where the energy consumption of the first sampling strategy is lower than that of the second sampling strategy. For example, the first sampling strategy includes low-frequency sampling, timed wake-up, local caching, and periodic summary uploading to reduce long-term operating energy consumption; the second sampling strategy includes high-frequency sampling, shortening the sampling period, improving analog-to-digital conversion resolution, adding upload fields, or enabling long-distance communication. Simultaneously, data before triggering is read from the circular buffer and encapsulated together with high-frequency data during triggering and recovery data after triggering into an event data packet. In some implementations, a switching threshold is also set, which is lower than a preset threshold. When the event confidence level is lower than the switching threshold for a preset duration, the system reverts to the first sampling strategy to form a hysteresis interval, reducing repeated switching near the threshold.

[0113] Secondly, when a monitoring point enters an event state, it sends a mutual trigger command to a preset number of neighboring monitoring points. After receiving the mutual trigger command, the neighboring monitoring points enter the verification sampling state. Even if their own event confidence has not yet reached the entry threshold, they increase the sampling density and return a feature summary in a short period of time. Furthermore, they are sorted according to the arrival time / node position of the feature summary response to form an event chain of the track segment.

[0114] In this embodiment, the feature summary includes node location, timestamp, event confidence level, vibration amplitude, compensation strain value, and correlation C. Node location can be obtained from the installation location table, sleeper number, or node configuration parameters. The track segment event chain is represented as: {node identifier, node location, response arrival time, event confidence level, vibration amplitude, compensation strain value, and correlation C}.

[0115] Furthermore, based on the relationship between the features in the feature summary of each event chain and the change of node position, the event type is determined: if the event response is concentrated on a single sleeper or a few adjacent sleepers, it is determined to be a local impact or a local support anomaly; if the event response propagates regularly along multiple sleepers, the impact propagation direction or train passing characteristics are determined by combining the trigger time sequence.

[0116] For example, when the highest event confidence and maximum vibration amplitude in the event chain are concentrated on a single sleeper or a few adjacent sleepers, and exhibit a decaying distribution towards both lateral nodes, the sleeper or adjacent sleeper section can be output as an impact location or a location of localized abnormal stress. When the compensation strain values ​​of multiple adjacent nodes deviate from the background baseline for a long period and the vibration impact characteristics are not significant, the corresponding section can be output as an abnormal support section. When the event response occurs sequentially along multiple sleepers in chronological order, and the vibration amplitude and compensation strain response conform to the propagation relationship in the direction of train passage, the impact propagation direction or train passage characteristics can be output.

[0117] To further optimize the above technical solution, the background baseline, compensation coefficient, and trigger parameters are updated based on the long-term event chain results, and the updated parameters are then sent to the corresponding monitoring devices. Thus, the system not only completes data aggregation but also forms a collaborative closed loop with the front-end nodes.

[0118] Specifically, during parameter updates, the aggregation node or host computer can statistically analyze the results of non-event windows and confirmed event chains within a preset period. For non-event windows, the background baseline and compensation coefficients are updated using moving average, exponential smoothing, or recursive least squares methods; for confirmed real event windows, they do not participate in the background baseline and compensation coefficient updates. For trigger parameters, the aggregation node or host computer can adjust the entry threshold, exit threshold, correlation threshold, background stability threshold, and mutual trigger node range based on the number of false triggers, missed trigger records, verification results of neighboring nodes, and event confidence distribution. The updated parameters can include a version number, applicable node identifier, and effective time, and are distributed to the corresponding monitoring device via the communication module; the monitoring device enables the new parameters after confirming that the parameter verification has passed, and temporarily suspends or freezes parameter updates during event states.

[0119] It should be noted that the step numbers above are for descriptive convenience only and do not constitute a strict limitation on the order of step execution. In some implementations, some steps can be executed in parallel or their order can be adjusted in an equivalent manner.

[0120] In one exemplary embodiment, the overall process of the collaboratively triggered compensation sleeper monitoring method includes:

[0121] Step 1: Obtain the strain response related to the force path under the track through the strain acquisition module, obtain the vibration response of the area under the track through the vibration acquisition module, and obtain the first temperature and external reference temperature and humidity through the first environmental unit and the second environmental unit.

[0122] Step 2: The main control module performs time alignment on all data and determines the equivalent temperature of the elastic measuring unit based on the first temperature, the external reference temperature and humidity, and the rate of change of both.

[0123] Step 3: The main control module uses a temperature and humidity coupling compensation model to compensate for the strain response and / or vibration characteristics, and obtains the strain compensation value and vibration compensation value.

[0124] Step 4: The main control module extracts the compensated multi-source features within the sliding time window, calculates the event confidence by combining it with the background baseline of the non-event window, and determines whether to enter or exit the event state based on the entry threshold, exit threshold and consistency condition.

[0125] Step 5: Perform low-power inspection sampling when not in the event state; perform high-frequency event sampling when in the event state, and encapsulate the pre-trigger cached data, the data during the trigger, and the post-trigger recovery data into an event data packet.

[0126] Step 6: The monitoring device that has entered the event state sends a mutual trigger command to the neighboring monitoring device. The neighboring monitoring device enters the verification sampling state and returns the feature summary and timestamp.

[0127] Step 7: The aggregation node or host computer forms an event chain for the track section based on the event confidence, response arrival time and spatial decay relationship of multiple monitoring devices, and outputs the diagnostic results of impact location, abnormal support section or local force anomaly.

[0128] In step 7, the aggregation node or host computer can first receive feature summaries uploaded by multiple monitoring devices and query the corresponding installation location based on the node identifier; then, the summary data within the same event time range are clustered and an event chain is formed according to the response arrival time and node location; then, the main response node is determined according to the event confidence level, the propagation order is determined according to the response arrival time difference, and the abnormal concentration area is determined according to the vibration amplitude attenuation and compensation strain distribution; finally, the diagnostic results including impact location, abnormal support section or local force abnormality type are output.

[0129] Any content or technical means not mentioned in the embodiments of this invention can be obtained by referring to the prior art. This disclosure does not limit the scope of the invention and therefore will not be elaborated further.

[0130] The embodiments of the present invention have been described in detail above, and the principles and implementation methods of the present invention have been explained. The descriptions of the embodiments above are only for the purpose of helping to understand the method and core ideas of the present invention. Those skilled in the art should understand that the embodiments of the present invention can be provided as methods, computer software program products, or electronic devices, etc. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present invention can take the form of a computer program product implemented on one or more storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0131] It should be noted that the word "comprising" does not exclude the presence of components or steps not listed in the claims. The words "a" or "an" preceding a component do not exclude the presence of a plurality of such components. This invention can be implemented by means of hardware comprising several different components and by means of a suitably programmed computer.

[0132] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on its differences from other embodiments. Similar or identical parts between embodiments can be referred to interchangeably. For the apparatus disclosed in the embodiments, since they correspond to the methods disclosed in the embodiments, the description is relatively simple; relevant parts can be referred to the method section.

[0133] The above description of the disclosed embodiments enables those skilled in the art to make or use the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the invention. Therefore, the invention is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A collaborative triggering compensation-type sleeper monitoring method, characterized in that, include: Real-time monitoring of strain and vibration signals under the sleepers, and acquisition of the first temperature of the monitoring area; The ambient temperature and humidity are acquired, and the strain signal and the vibration signal are compensated based on the ambient temperature and humidity and the first temperature to obtain the strain compensation signal and the vibration compensation signal. Multiple features are extracted based on the strain compensation signal and vibration compensation signal, and the event confidence is calculated using the features; The event confidence level is used to trigger the construction of the track segment event chain, and the event type is determined based on the track segment event chain.

2. The collaborative triggering compensation sleeper monitoring method as described in claim 1, characterized in that, Compensation for the strain signal based on the external environmental temperature and humidity and the first temperature includes: ； In the formula, Indicates strain compensation signal, This represents the strain signal monitored in real time. Indicates the strain signal of the reference channel. Indicates the reference channel compensation coefficient. Indicates the equivalent temperature of the monitored area. Indicates the reference temperature of the monitoring area. This represents the temperature difference between the initial temperature and the ambient temperature. Indicates the humidity of the external environment. This represents the strain compensation coefficient.

3. The collaborative triggering compensation sleeper monitoring method as described in claim 1, characterized in that, Compensating the vibration signal based on the external environmental temperature and humidity and the first temperature includes: ； In the formula, Indicates vibration compensation signal, This indicates the vibration signal being monitored in real time. The equivalent temperature of the monitoring area Indicates the reference temperature of the monitoring area. Indicates the humidity of the external environment. This represents the vibration compensation coefficient.

4. The collaborative triggering compensation sleeper monitoring method as described in claim 2 or 3, characterized in that, The equivalent temperature of the monitored area is obtained as follows: ； In the formula, Represents the smoothing coefficient. Indicates the first temperature. Indicates the ambient temperature; or ； In the formula, t represents the current sampling time. Indicates based on the first temperature relative to external ambient temperature The temperature lag time determined by the rate of change or historical correlation, This indicates the effective temperature sampling time after considering thermal inertia hysteresis.

5. The collaborative triggering compensation sleeper monitoring method as described in claim 1, characterized in that, The various features are extracted based on a sliding time window, including strain-vibration time correlation, and Strain peak value and / or strain rate of change extracted from strain compensation signal; The vibration peak value, root mean square value, rate of change of acceleration, impact kurtosis and / or frequency band energy are extracted from the vibration compensation signal.

6. The collaborative triggering compensation sleeper monitoring method as described in claim 1, characterized in that, Calculating event confidence using the aforementioned features includes: Determine the mean and standard deviation of the background baseline corresponding to each feature; The current feature is normalized based on the mean and standard deviation. The normalized features are weighted and summed.

7. The collaborative triggering compensation sleeper monitoring method as described in claim 1, characterized in that, The event chain is constructed based on the event confidence level, and the event type is determined based on the event chain, including: When the confidence level of the event is higher than a preset threshold, a mutual trigger command is sent to a preset number of neighboring monitoring points. Receive the feature summaries returned by the neighboring monitoring points in response to the mutual triggering command, and sort them according to the response arrival time and node location to form a track segment event chain; The event type is determined based on the relationship between the changes in features and node positions in the feature summaries of the event chain of the track segment.

8. The collaborative triggering compensation sleeper monitoring method as described in claim 7, characterized in that, When the confidence level of an event is lower than a preset threshold, a first sampling strategy is used for monitoring; otherwise, a second sampling strategy is used for monitoring, wherein the energy consumption of the first sampling strategy is lower than that of the second sampling strategy.

9. A collaborative triggering compensation type sleeper monitoring device, characterized in that, The method for implementing the collaborative triggering compensation sleeper monitoring method according to any one of claims 1-8 includes: an elastic measuring unit and a force coupling seat; The elastic measurement unit is fixed above the force coupling seat and includes an environmental reference cavity and a sealed acquisition cavity inside; a strain acquisition module and a vibration acquisition module are provided on the top. The strain acquisition module is used to acquire the strain response transmitted to the elastic measurement unit via the force coupling seat; The vibration acquisition module is used to acquire the vibration response transmitted to the elastic measuring unit via the force coupling seat; The environmental acquisition chamber includes a first environmental unit and a second environmental unit; the first environmental unit is used to acquire a first temperature of the monitoring area, and the second environmental unit is used to acquire the external environmental temperature and humidity. The sealed acquisition cavity is equipped with a main control module, which is used to compensate the strain signal and the vibration signal based on the external ambient temperature and humidity and the first temperature, respectively, to obtain strain compensation signal and vibration compensation signal; extract various features based on the strain compensation signal and vibration compensation signal, and use the features to calculate event confidence; and trigger the construction of track segment event chain according to the event confidence, and determine the event type according to the track segment event chain.

10. A collaborative triggering compensation sleeper monitoring system, characterized in that, It includes multiple collaborative triggering compensation sleeper monitoring devices as described in claim 9, and a host computer; The host computer is used to receive feature summaries uploaded by each collaboratively triggered compensating sleeper monitoring device to construct a track section event chain, and to determine the event type based on the track section event chain.

Images