Railway vibration isolation elastic element condition detection system and method

Abstract

The application provides a track vibration isolation elastic element state detection system and method, the system comprises a plurality of vibration detection devices containing vibration acceleration sensors, an excitation device containing force sensors and a calculation and analysis device, and the plurality of vibration detection devices are respectively arranged at each predetermined measuring point in the floating slab track, each measuring point is located on one side of one of the vibration isolators, the measuring points are divided into key measuring points and adjacent measuring points, and the excitation point is arranged on the lower foundation on one side of the key measuring point, so that the excitation point can be excited by the excitation device and force signal data can be obtained, and the vibration response to the excitation can be obtained by the plurality of vibration detection devices, the vibration response can be analyzed in combination with the distribution and correlation of the measuring points, the relative position relationship between the measuring points and the excitation point and the principle of the vibration response, and the state of the elastic element of the vibration isolator in the floating slab track bed can be determined, so that the problem of steel spring fracture can be found in time, and the operation safety of the floating slab track is ensured.

Description

Technical Field

[0001] This invention belongs to the field of rail transit facility condition monitoring technology, specifically relating to a condition monitoring system and method for track vibration damping elastic elements. Background Technology

[0002] Currently, floating slab track beds are used in track systems with special vibration reduction requirements. These typically consist of multiple track slabs and several vibration-isolation elastic elements. Each track slab is supported on the underlying foundation by multiple elastic elements. When a train passes, the elastic deformation of these elements absorbs vibration energy, reducing the amount of energy transmitted to the underlying foundation and achieving good vibration isolation. Currently used vibration-isolation elastic elements include rubber springs, steel springs, and combined springs. These elastic elements function as vibration damping components and key load-bearing parts. Damage to these elements (such as a broken steel spring) can seriously affect train operation safety; therefore, the inspection of elastic elements is crucial. Traditionally, each elastic element is inspected manually, using visual inspection or mechanical tools. This method is not only inefficient, leading to long downtime for maintenance, but also makes it difficult to detect damage such as a broken steel spring in a timely manner, thus creating safety hazards. Replacing all elastic elements in bulk prematurely to eliminate safety hazards results in considerable waste and high maintenance costs.

[0003] Therefore, in order to improve the safety of floating slab track, increase inspection efficiency, and reduce downtime for maintenance, there is an urgent need for a fast, accurate, and non-destructive method for detecting the condition of elastic elements. Summary of the Invention

[0004] This invention addresses the aforementioned problems by providing a rapid, accurate, and non-destructive method and system for detecting the state of elastic elements in floating slab tracks. The invention employs the following technical solution:

[0005] This invention provides a track vibration isolation elastic element state detection system, installed in a floating slab track having a subbase, track bed slab, and multiple vibration isolators, for detecting the state of the elastic elements in the vibration isolators. The system comprises: multiple vibration detection devices, each installed at predetermined measuring points in the floating slab track, including vibration acceleration sensors for acquiring vibration signal data; an excitation device, including force sensors, for exciting predetermined excitation points in the floating slab track and acquiring corresponding force signal data; and a calculation and analysis device for analyzing the vibration response to the excitation based on the force signal data and the vibration signal data at each measuring point, and determining the state of the elastic element based on the analysis results. The measuring points include at least one key measuring point and one adjacent measuring point. The key measuring point is located on the upper surface of the track bed slab on one side of one of the vibration isolators, and the adjacent measuring point is located on the upper surface of the track bed slab on one side of another vibration isolator adjacent to the key measuring point. The excitation point is located on the subbase on the side of the vibration isolator corresponding to the key measuring point.

[0006] The track vibration isolation elastic element state detection system provided by this invention may also have the following technical features: the system further includes a signal acquisition device, connected to the vibration acceleration sensor, the force sensor, and the calculation and analysis device, for acquiring the vibration signal data and the force signal data and transmitting them to the calculation and analysis device. The vibration detection device further includes a metal counterweight, disposed at the measuring point; and an insulating magnetic base, adsorbed and fixed on the counterweight, with the vibration acceleration sensor fixed to the counterweight via the insulating magnetic base. The excitation device is a force hammer with the force sensor built-in.

[0007] The track vibration isolation elastic element state detection system provided by this invention may also have the following technical features: The number of adjacent measuring points is one or two. Each measuring point is located on the side of the corresponding vibration isolator near the center of the track bed slab. The horizontal direction connecting the measuring point to the center of the corresponding vibration isolator, and the horizontal direction connecting the key measuring point to the excitation point, are both along the width direction of the floating slab track. The horizontal distance between the measuring point and the center of the corresponding vibration isolator is 15cm to 25cm. The lower foundation is a tunnel wall, and the vertical distance between the excitation point and the key measuring point is 1m to 1.8m.

[0008] The track vibration isolation elastic element state detection system provided by the present invention may also have the following technical features, wherein the calculation and analysis device includes: a vibration signal acquisition unit for acquiring corresponding vibration signal data from each of the vibration acceleration sensors; a force signal acquisition unit for acquiring force signal data from the force sensors; a data preprocessing unit for preprocessing the vibration signal data and the force signal data; a state parameter calculation unit for calculating various state parameters based on the preprocessed vibration signal data and the force signal data; and an elastic element state determination unit for determining the state of the elastic element according to the state parameters and predetermined determination rules.

[0009] The track vibration isolation elastic element state detection system provided by this invention may also have the following technical features, wherein the state parameters include at least the response time of each measuring point to the excitation, the energy ratio of the force signal data and the vibration signal data, the attenuation rate of the vibration signal, and the natural frequency of the track bed slab containing the vibration isolator. The judgment rules include: when the main judgment criterion is met and at least one auxiliary judgment criterion is met, or when the main judgment criterion is not met but more than a predetermined number of auxiliary judgment criteria are met, the elastic element corresponding to the key measuring point is determined to be in an abnormal state. The main judgment criterion is: the response time of the key measuring point and the response time of the adjacent measuring points do not meet a predetermined response order. Multiple auxiliary judgment criteria include: the energy ratio of the key measuring point and the energy ratio of the adjacent measuring points do not meet a predetermined energy ratio relationship; the attenuation rate deviates from a predetermined attenuation rate reference value by more than a predetermined threshold; and the natural frequency deviates from a predetermined natural frequency reference value by more than a predetermined threshold.

[0010] The track vibration isolation elastic element state detection system provided by this invention may also have the following technical features: the response sequence is such that the response time of all adjacent measuring points is greater than the response time of the key measuring point; and the energy ratio relationship is such that the relative energy ratio of all adjacent measuring points is less than the relative energy ratio of the key measuring point.

[0011] The track vibration isolation elastic element state detection system provided by the present invention may also have the following technical features, wherein the calculation and analysis device includes: a vibration signal acquisition unit for acquiring corresponding vibration signal data from each of the vibration acceleration sensors; a force signal acquisition unit for acquiring force signal data from the force sensors; a data preprocessing unit for preprocessing the vibration signal data and the force signal data; a feature extraction unit for extracting features from the preprocessed vibration signal data and the force signal data to obtain multiple feature parameters that can reflect the structural dynamic characteristics of the floating slab track bed; a feature fusion unit for fusing the multiple feature parameters and combining the distribution and correlation information of multiple measuring points to obtain a fused feature vector; and an elastic element state determination unit for processing the fused feature vector and determining the state of the elastic element through a trained machine learning model.

[0012] The track vibration isolation elastic element state detection system provided by the present invention may also have the following technical features: the feature extraction unit extracts feature parameters in the time domain, frequency domain, and time-frequency domain; the feature parameters in the time domain include peak acceleration, root mean square value, crest factor, decay time, peak value and time delay of cross-correlation function; the feature parameters in the frequency domain include dominant frequency, peak amplitude, frequency band energy, frequency component complexity, transfer function amplitude / phase, and coherence function; the feature parameters in the time-frequency domain include fast Fourier transform energy features; the feature fusion unit integrates the feature parameters in the time domain, frequency domain, and time-frequency domain, and combines them with the distribution and correlation information of the measurement points to obtain a fused feature vector; the machine learning model is a hybrid model of convolutional neural network and bidirectional long short-term memory network, used to output a classification result of the elastic element state based on the input fused feature vector.

[0013] The track vibration isolation elastic element state detection system provided by the present invention may also have the following technical features: multiple vibration isolators corresponding to a track bed slab are all set to a normal state, and several of the multiple vibration isolators corresponding to a track bed slab are set to an abnormal state. The key measuring point and the excitation point are set on one side of each vibration isolator in sequence, and the vibration signal data and the force signal data are collected respectively and the corresponding label information is set to be used as training data to train the machine learning model.

[0014] This invention provides a method for detecting the state of track vibration isolation elastic elements, which has the following technical features: Step S1, setting up a track vibration isolation elastic element state detection system as described above in a floating slab track having a subbase, track bed slab, and multiple vibration isolators, wherein the vibration isolators include elastic elements; Step S2, using the excitation device to excite at the excitation point and acquiring force signal data during excitation; Step S3, collecting vibration signal data at each of the measuring points through multiple vibration detection devices; Step S4, analyzing the vibration response to the excitation based on the force signal data and the vibration signal data at each of the measuring points, and determining the state of the elastic element based on the analysis results.

[0015] The role and effect of invention

[0016] The track vibration isolation elastic element state detection system and method provided by the present invention includes multiple vibration detection devices containing vibration acceleration sensors, excitation devices containing force sensors, and calculation and analysis devices. The multiple vibration detection devices are respectively installed at predetermined measuring points in the floating slab track. Each measuring point is located on one side of one of the vibration isolators in the floating slab track bed. The measuring points are divided into key measuring points and adjacent measuring points. The excitation point is located on the lower foundation on one side of the key measuring point. Therefore, excitation (instantaneous force) can be applied to the excitation point through the excitation device and the corresponding force signal data can be obtained. The vibration response to the excitation can be obtained through multiple vibration detection devices. By combining the distribution, correlation, relative positional relationship between the measuring points and the excitation point, and the principle of vibration response, the vibration response can be analyzed to determine the state of the elastic element of the vibration isolator in the floating slab track bed. This allows for timely detection of problems such as steel spring breakage failure, ensuring the operational safety of the floating slab track. Furthermore, the system and method only require the installation of vibration detection devices on the upper surface of the track bed slab and excitation on the lower foundation; it does not require lifting the track bed slab or disassembling and assembling vibration isolators, making the detection very convenient and efficient. Attached Figure Description

[0017] Figure 1 This is a structural block diagram of the track vibration isolation elastic element state detection system in Embodiment 1 of the present invention;

[0018] Figure 2 This is a schematic diagram of the vibration detection device in Embodiment 1 of the present invention;

[0019] Figure 3 This is a schematic diagram of the distribution of predetermined points in the track in Embodiment 1 of the present invention. Figure 1 ;

[0020] Figure 4 This is a schematic diagram of the distribution of predetermined points in the track in Embodiment 1 of the present invention. Figure 2 ;

[0021] Figure 5 This is a schematic diagram of the distribution of predetermined points in the track in Embodiment 1 of the present invention. Figure 3 ;

[0022] Figure 6 This is a structural block diagram of the computational analysis device in Embodiment 1 of the present invention;

[0023] Figure 7 This is a flowchart of the method for detecting the state of track vibration isolation elastic elements in Embodiment 1 of the present invention;

[0024] Figure 8 This is a structural block diagram of the computational analysis device in Embodiment 2 of the present invention;

[0025] Figure 9 This is a flowchart of the method for detecting the state of the track vibration isolation elastic element in Embodiment 2 of the present invention.

[0026] Figure label:

[0027] Track vibration isolation elastic element state detection system 100; vibration detection device 110; vibration acceleration sensor 111; counterweight 112; insulating magnetic base 113; signal acquisition device 120; excitation device 130; force sensor 131; calculation and analysis device 140; data storage unit 1401; vibration signal acquisition unit 1402; force signal acquisition unit 1403; metadata acquisition unit 1404; data preprocessing unit 1405; state parameter calculation unit 1406; elastic element state determination unit 1407; analysis side control unit 1408; feature extraction unit 1409; feature fusion unit 1410; floating slab track bed 200; substructure 210; track bed slab 220; vibration isolator 230; rail 240; excitation point A1; adjacent measuring points C1, C3; key measuring point C2; vibration isolator center O1~O8. Detailed Implementation

[0028] To make the technical means, creative features, objectives and effects of the present invention easy to understand, the following describes in detail the track vibration isolation elastic element state detection method and system of the present invention with reference to embodiments and accompanying drawings.

[0029] <Example 1>

[0030] This embodiment provides a method and system for detecting the state of track vibration isolation elastic elements. The system composition will be described first, and then the specific detection method will be explained in conjunction with the system composition.

[0031] Figure 1 This is a structural block diagram of the track vibration isolation elastic element state detection system in this embodiment.

[0032] like Figure 1As shown, the track vibration isolation elastic element condition detection system 100 includes multiple vibration detection devices 110, one or more signal acquisition devices 120, one or more excitation devices 130, and a calculation and analysis device 140.

[0033] The vibration detection device 110 includes at least a vibration acceleration sensor, which is installed in the actual floating slab track to detect the vibration acceleration signal related to the vibration isolation elastic element of the floating slab.

[0034] Figure 2 This is a schematic diagram of the vibration detection device in this embodiment.

[0035] like Figure 2 As shown, in this embodiment, the vibration detection device 110 includes a vibration acceleration sensor 111, a counterweight 112, and an insulating magnetic base 113.

[0036] Among them, the vibration acceleration sensor 111 is a triaxial vibration acceleration sensor that can measure vibration acceleration data with a range of ±10g.

[0037] The counterweight 112 is used to set in the floating slab track and to place the vibration acceleration sensor 111, so that the vibration acceleration sensor 111 can remain stable during the test. In this embodiment, the counterweight 112 is a rectangular metal counterweight that can be magnetically attracted, and its dimensions are 100mm*50mm*10mm (length*width*height).

[0038] The insulating magnetic base 113 is cylindrical in shape and has a mounting hole in the middle for fixing the vibration acceleration sensor 111 to the counterweight 112 by magnetic attraction.

[0039] During use, the insulating magnetic base 113 is firmly attached to the designated mounting point on the surface of the counterweight 112 to ensure reliable mechanical coupling. During installation, pay attention to the installation direction to ensure that the sensitive axis (usually the Z-axis) of the vibration acceleration sensor 111 is vertically upward or vertically downward to effectively measure vertical vibration.

[0040] The signal acquisition device 120 is connected to multiple vibration detection devices 110 via cables to acquire vibration acceleration signal data detected by the sensors of these vibration detection devices 110. In this embodiment, the signal acquisition device 120 is model INV3062V, and its measurement range is ±10V.

[0041] The excitation device 130 applies a certain excitation to the lower foundation, causing it to vibrate. This vibration is transmitted to multiple vibration-damping elastic elements in contact with the lower foundation, allowing for state monitoring of these elements. In this embodiment, the excitation device 130 is a hammer used to excite the lower foundation by striking it. Preferably, at least the hammerhead is made of rubber. Preferably, the hammer has a built-in force sensor 131 that records the striking force pulses, facilitating the control of the consistency of energy input across multiple strikes. The signal acquisition device 120 is also connected to the force sensor 131 via a cable to acquire the force signal data measured by the force sensor 131.

[0042] In an alternative, the excitation device 130 can also be an automated device. For example, the excitation device 130 includes a hammer with a built-in force sensor, a rotating mechanism, and a fixing mechanism. The fixing mechanism is used to temporarily fix the excitation device 130 at a specific position on the lower foundation. The rotating mechanism is mounted on the fixing mechanism and is used to drive the hammer to strike the excitation point on the lower foundation with a predetermined force.

[0043] The calculation and analysis device 140 is connected to the signal acquisition device 120 via cable or wireless communication. It is used to acquire the signals acquired by the signal acquisition device 120 and perform calculation and analysis based on these signals to obtain the state information of the elastic element, etc.

[0044] The following will first describe the specific layout of the sensors and the corresponding signals collected, and then, in conjunction with the signals, explain the functions implemented by the calculation and analysis device 140.

[0045] Figure 3 This is a schematic diagram of the distribution of predetermined points in the track in this embodiment. Figure 1 (Top view of the track) Figure 4 This is a schematic diagram of the distribution of predetermined points in the track in this embodiment. Figure 2 (Top view of the track bed slab). Figure 5 This is a schematic diagram of the distribution of predetermined points in the track in this embodiment. Figure 3 (Schematic diagram of track cross section).

[0046] like Figures 3 to 5As shown, the floating slab track 200 includes a lower foundation 210, multiple track slabs 220 arranged sequentially, multiple vibration isolators 230, and steel rails 240. The lower foundation 210 can be a flat surface or a tunnel wall, etc. The track slab 220 is a monolithic concrete track slab with multiple vibration isolator mounting holes extending along its thickness. These mounting holes are, for example, six or eight equally spaced holes arranged in two rows along the length of the track slab 220. Each vibration isolator mounting hole contains one vibration isolator 230, and the track slab 220 is floating on the lower foundation 210 via the multiple vibration isolators 230. Two steel rails 240 are mounted parallel to each other on the track slab 220 and are located above or to the side of the two rows of vibration isolators 230, respectively. For ease of description below, the centers of the eight vibration isolators 230 in one track slab 220 in this embodiment are designated O1 to O8 from a top-view perspective.

[0047] Each vibration isolator 230 includes, for example, an outer sleeve embedded in the track slab 220 and an elastic unit disposed at the lower part of the outer sleeve. The inner wall of the outer sleeve has a ring of load-bearing protrusions that match the elastic unit. The upper end of the elastic unit abuts against the ring of load-bearing protrusions, and the lower end rests on the lower foundation. The elastic unit contains an elastic element, which can be, for example, a helical steel spring or a rubber spring. The structure of the vibration isolator 230 is prior art and will not be described in detail.

[0048] The excitation device 140 performs excitation at a predetermined excitation point on the lower foundation. The excitation point is located next to one of the vibration isolators 230, and the horizontal line connecting the excitation point and the center of the vibration isolator 230 (the central axis of its outer sleeve) is in the direction of the track width. In this embodiment, the excitation point A1 is located on the tunnel wall next to one of the vibration isolators 230 in the middle of the track bed slab 220, and is located above the vibration isolator 230. The vertical distance h1 between the excitation point A1 and the upper surface of the track bed slab 220 can be 1m to 1.8m, preferably 1.5m. During testing, the excitation device 140 (hammer) is used to perform a momentary vertical strike (perpendicular to the surface direction of the tunnel wall at the excitation point A1) on the excitation point A1 to excite it. The strike should be crisp and fast to ensure consistent energy input.

[0049] Multiple vibration detection devices 110 are installed at multiple predetermined test points (hereinafter referred to as test points) in the track. All test points are located on the upper surface of the track slab 220 and inside the rail. Each test point is located next to a vibration isolator 230, and the horizontal line connecting the center of the test point and the corresponding vibration isolator 230 is in the width direction of the track. The multiple test points include at least one key test point and one or two adjacent test points. The key test point is the test point closest to the excitation point among the multiple test points. The vibration isolators 230 corresponding to the adjacent test points and the key test point are two adjacent vibration isolators 230.

[0050] In this embodiment, each track slab 220 is equipped with eight vibration isolators 230, and a total of three measuring points are provided: a key measuring point C2 and two adjacent measuring points C1 and C3, which are respectively set next to the three vibration isolators 230 arranged in sequence. The two adjacent measuring points C1 and C3 are symmetrically set on both sides of the key measuring point C2. The three measuring points C1 to C3 are distributed on the same straight line and are all located on the side of the corresponding vibration isolator closer to the middle of the track slab. The horizontal distance d1 between each measuring point and the center of the corresponding vibration isolator 230 (i.e., the center O6 of the vibration isolator for the key measuring point C2, the center O4 of the vibration isolator for the adjacent measuring point C1, and the center O8 of the vibration isolator for the adjacent measuring point C3) is 15cm to 25cm, preferably 20cm. In an alternative embodiment, only the key measuring point C2 and the adjacent measuring point C1, or only the key measuring point C2 and the adjacent measuring point C3, can be provided. For example, when the key measuring point C2 is set on one side of a vibration isolator 230 at one end of the track slab 220, there is only one adjacent measuring point.

[0051] With this setting of measuring points, the excitation generated at the excitation measuring point A1 can be transmitted more evenly to the three measuring points C1~C3 at the same time. It also facilitates the synchronous response comparison analysis between the key measuring point C2 and the adjacent measuring points C1 or C3, thereby reflecting the state of the support structure (i.e., multiple vibration isolators 230 in a track bed slab 220).

[0052] Figure 6 This is a structural block diagram of the computational analysis device in this embodiment.

[0053] like Figure 6 As shown, the calculation and analysis device 140 includes a data storage unit 1401, a vibration signal acquisition unit 1402, a force signal acquisition unit 1403, a metadata acquisition unit 1404, a data preprocessing unit 1405, a state parameter calculation unit 1406, an elastic element state determination unit 1407, an analysis-side control unit 1408, an input display unit 1411, and an analysis-side communication unit 1412.

[0054] The data storage unit 1401 stores various predetermined detection parameters and calculation / analysis parameters, and also stores acquired vibration acceleration signals, calculation / analysis results, and determined vibration isolation elastic element status information. Detection parameters include the size of the acquisition time window, etc.

[0055] The vibration signal acquisition unit 1402 is used to acquire vibration acceleration signals at different points measured by the vibration acceleration sensors 111 of the multiple vibration detection devices 110 through the signal acquisition device 120. In this embodiment, vibration acceleration signal data at the above three measuring points are acquired as three-channel signal data. The vibration acceleration signal data is raw waveform data including timestamps. Furthermore, the predetermined acquisition time window size ensures that the acquired three-channel signals include the complete impact response decay process, thereby ensuring that the three-channel signals can cover the main modal responses.

[0056] The force signal acquisition unit 1403 is used to acquire force signal data of the force measured by the force sensor 131 when the excitation device 130 is excited by the signal acquisition device 120. Similarly, the predetermined acquisition time window size ensures that the acquired force signal data includes the complete impact response decay process.

[0057] The metadata acquisition unit 1404 is used to acquire other metadata required for analysis. In this embodiment, the metadata acquisition unit 1404 is used to provide a corresponding interactive interface for testers to input metadata such as the test number, time, tester information, and environmental condition information. Environmental condition information may include, for example, the temperature and humidity of the environment during the test, and the ambient noise.

[0058] The data preprocessing unit 1405 is used to preprocess the acquired three-channel signal data and force signal data to improve the quality of the signal data. Preprocessing may include trend term removal, filtering, windowing, etc.

[0059] The state parameter calculation unit 1406 calculates various state parameters based on the preprocessed three-channel signal data and force signal data.

[0060] In this embodiment, the calculated state parameters include at least the response time of each measuring point to the excitation, the energy ratio corresponding to the force signal and the vibration signal, the attenuation rate of the vibration signal, and the natural frequency of the floating slab track bed. These state parameters can all be calculated using methods in the prior art. For example, the response time can be calculated based on the peak time of the force signal data and the peak time of the vibration signal data of each measuring point; by integrating the force signal data and the vibration signal data respectively, the force signal energy and vibration energy can be obtained, and these can be used as the input energy and output energy respectively to calculate the energy ratio.

[0061] The elastic element state determination unit 1407 is used to determine the state of the elastic elements in the support structure (outer sleeve of the vibration isolator and elastic element area) of the floating slab track bed based on multiple state parameters and predetermined determination rules. In this embodiment, the determination rules include: when the main determination criteria are met and at least one auxiliary determination criterion is met, or when the main determination criteria are not met but more than a predetermined number (e.g., more than two) of the auxiliary determination criteria are met, the elastic element in the vibration isolator corresponding to the key measuring point is determined to be in an abnormal state. When the main determination criteria are met but all auxiliary determination criteria are not met, the elastic element corresponding to the key measuring point is determined to be in an undetermined state.

[0062] The main criterion is that the response time of the key measuring point and the response time of the adjacent measuring points do not meet the predetermined response sequence.

[0063] When the elastic element in the vibration isolator 230 (centered at O6) corresponding to the critical measuring point C2 is in normal condition (i.e., no spring breakage when it is a steel spring), the vibration will be transmitted to this elastic element first because it is closest to the excitation point A1. The response time (i.e., the energy transmission time of the impact excitation) of the vibration acceleration sensor 111 located at the critical measuring point C2 should be the shortest, while the response times of adjacent measuring points should all be longer than the response time of the critical measuring point C2. However, if the elastic element corresponding to the critical measuring point C2 is in an abnormal state (e.g., a broken steel spring), the vibration will bypass this elastic element and be transmitted to the elastic element of the adjacent vibration isolator 230, causing the response time of the adjacent measuring points C1 / C3 to be earlier than the response time of the critical measuring point C2.

[0064] Multiple auxiliary judgment criteria include: the natural frequency of the track slab containing the vibration isolator deviates from the predetermined natural frequency reference value, and the deviation value exceeds the predetermined threshold; the energy ratio of key measuring points and the relationship between the energy ratio of adjacent measuring points do not meet the predetermined energy ratio relationship; the attenuation rate of the vibration signal deviates from the predetermined attenuation rate reference value, and the deviation value exceeds the predetermined threshold.

[0065] Regarding the natural frequency offset, the natural frequency of the floating slab track bed can be analyzed using the excitation response method based on the collected vibration acceleration signal data. The predetermined natural frequency reference value can be calculated according to the following formula:

[0066]

[0067] In the formula, k The total stiffness of multiple elastic elements, m This refers to the total mass of the track slab including the vibration isolators. In this embodiment, the reference value for the natural frequency is 11Hz ± 5%.

[0068] When an elastic element is in an abnormal state (e.g., a broken steel spring), the natural frequency of the floating slab track bed as a whole / partially will change due to stiffness. k The natural frequency decreases as the measured data decreases. Therefore, if the natural frequency calculated from the measured data differs significantly from the predetermined natural frequency reference value, it can be used as an auxiliary basis for judging one or more of the multiple elastic elements in an abnormal state.

[0069] Regarding the energy ratio, since the key measuring point C2 is closest to the excitation point A1, the vibration acceleration sensor 111 located at the key measuring point C2 should receive the maximum energy. The energy ratio (the ratio of output energy to input energy) is E. r =E out / Ein The energy ratio of neighboring measuring points C1 / C3 should be the largest, and the energy ratio of the key measuring point C2 should be less than that of the key measuring point C2. If the energy ratio of neighboring measuring points C1 / C3 is greater than that of the key measuring point C2, and the difference between the energy ratios of neighboring measuring points and the key measuring point exceeds a predetermined threshold, it can be used as an auxiliary basis for judging the abnormal state of the elastic unit corresponding to the key measuring point C2.

[0070] The attenuation rate benchmark value can be obtained, for example, by taking the average value of the vibration signal attenuation rate of the elastic element in normal state in the same type of floating plate track through multiple tests. The predetermined threshold can be set, for example, ±5% to ±20% of the benchmark value.

[0071] Alternatively, time-domain waveform difference features, such as waveform duration and number of peaks, can be extracted from a large number of experimental samples for classification and identification, thereby assisting in the determination of the state of the elastic element.

[0072] Alternatively, different weight values ​​can be set for multiple auxiliary judgment criteria, and a score can be calculated based on the number of qualified auxiliary judgment criteria and their weight values. When the main judgment criteria are not qualified, the state of the elastic element is determined based on the score and a predetermined score threshold.

[0073] The input display unit 1411 can be used by testers to set parameters, start data calculation and analysis processes, etc., and can be used to display raw signal data, intermediate data (preprocessed signal data, calculated state parameters, etc.), and determined state information. The aforementioned metadata acquisition unit 1404 and the input display unit 1411 can be the same functional unit.

[0074] The analysis-side communication unit 1412 is used to communicate with other devices, such as other computing and analysis devices, to transmit raw signal data and determined status information to other computing and analysis devices for further analysis, or to communicate with terminals held by testers and managers, to send status information to the terminals, etc.

[0075] The analysis-side control unit 1408 is used to control the operation of the above-mentioned functional units, including controlling the data storage unit 1401 to store the original signal data after the vibration signal acquisition unit 1402 and the force signal acquisition unit 1403 acquire the corresponding vibration acceleration signal data and force signal data; sequentially controlling the data preprocessing unit 1405, the state parameter calculation unit 1406, and the elastic element state determination unit 1407 to perform data preprocessing, calculation of various state parameters, and state determination of the elastic element, and controlling the data storage unit 1401 to store the preprocessed data, the calculated state parameters, and the determined state information during the process; and controlling the input display unit 1411 to display the intermediate data, state information, etc. accordingly.

[0076] Figure 7 This is a flowchart of the method for detecting the state of the track vibration isolation elastic element in this embodiment.

[0077] like Figure 7 As shown, based on the above-mentioned track vibration isolation elastic element state detection system 100, in this embodiment, the corresponding detection method includes the following steps:

[0078] Step S1: Install a track vibration isolation elastic element status detection system in the floating plate track.

[0079] Step S2: Use the excitation device to excite at the excitation point in the floating plate track and acquire the force signal data during excitation.

[0080] Step S3: Vibration signal data at each test point is collected using multiple vibration detection devices.

[0081] Step S4: Analyze the force signal data and vibration signal data of each measuring point, and determine the state of the vibration isolation elastic element based on the analysis results.

[0082] The steps described above will be explained in detail below.

[0083] Step S1: Install a track vibration isolation elastic element status detection system in the floating plate track.

[0084] As described above, a vibration isolator 230 is selected in the middle of the track bed slab 220. A vibration detection device 110 is installed at a key measuring point C2 on one side of the vibration isolator 230. Vibration detection devices 110 are also installed at adjacent measuring points C2 and / or C3. Each vibration acceleration sensor 111 is connected to a signal acquisition device 120 via cables. The signal acquisition device 120 is connected to a calculation and analysis device 140. During installation, a counterweight 112 is arranged at the measuring point. After the vibration acceleration sensor 111 is mounted on an insulating magnetic base 113, it is adsorbed and fixed to the upper surface of the counterweight 112, and the sensitive axis (Z-axis) of the vibration acceleration sensor 111 is vertically upward or vertically downward.

[0085] Before the formal test begins, the equipment should be debugged and its normal operation observed. Furthermore, to ensure data reliability and reproducibility, effective vertical tapping can be performed 3-5 times at excitation point A1, and signal data should be collected each time. After each tap, the validity of the signal data should be checked, for example, to see if there is significant saturation, noise contamination, or tapping failure. After multiple taps, the consistency of the collected signal data should be checked. If the measured signal data is problematic, or if there is significant inconsistency among multiple collected signal data, the equipment should be inspected and adjusted.

[0086] In addition, before the test begins and throughout the data acquisition process, ensure that there are no significant sources of interference or vibration near the test site, such as the operation of large equipment, personnel movement, vehicle traffic, or other impacts. If necessary, background noise and vibration measurements can be taken and recorded. In subsequent signal data processing, the measured background noise and vibration can be used to preprocess the signal data to minimize their influence.

[0087] Step S2: Use the excitation device to excite at the excitation point in the floating plate track and acquire the force signal data during excitation.

[0088] The device uses an excitation device 130 (force hammer) to strike the excitation point A1 on one side of the key measuring point C2. The force signal acquisition unit 1403 acquires the force signal data measured by the force sensor 131 through the signal acquisition device 120.

[0089] Step S3: Vibration signal data at each measuring point is collected using multiple vibration detection devices.

[0090] The vibration signal acquisition unit 1402 acquires vibration acceleration signal data measured by three vibration acceleration sensors 111 at three measuring points through the signal acquisition device 120, and uses it as three-channel signal data. Optionally, the tester can input other required metadata.

[0091] Step S4: Analyze the force signal data and vibration signal data of each measuring point, and determine the state of the vibration isolation elastic element based on the analysis results.

[0092] like Figure 7 As shown, step S4 specifically includes the following sub-steps:

[0093] Step S4-1 involves preprocessing the force signal data and the vibration signal data at each measuring point. This is the preprocessing performed by the data preprocessing unit 1405 mentioned above.

[0094] Step S4-2: Based on the preprocessed force signal data and the vibration signal data of each measuring point, various state parameters are calculated. This is the calculation performed by the state parameter calculation unit 1406 mentioned above.

[0095] Step S4-3: Based on the state parameters and predetermined judgment rules, the state of the vibration isolation elastic element is determined and the corresponding state information is output. That is, the judgment performed by the elastic element state determination unit 1407 mainly determines the state of the elastic element corresponding to the key measuring point C2.

[0096] Functions and effects of Example 1

[0097] According to the method and system for detecting the state of elastic elements of track vibration isolation provided in this embodiment, the system includes multiple vibration detection devices containing vibration acceleration sensors, excitation devices containing force sensors, and calculation and analysis devices. The multiple vibration detection devices are respectively installed at predetermined measuring points in the floating slab track. Each measuring point is located on one side of one of the vibration isolators in the floating slab track bed. The measuring points are divided into key measuring points and adjacent measuring points. The excitation point is set on the lower foundation on one side of the key measuring point. Therefore, excitation (instantaneous force) can be applied to the excitation point through the excitation device and the corresponding force signal data can be obtained. The vibration response to the excitation can be obtained through multiple vibration detection devices. By combining the distribution, correlation, relative positional relationship between the measuring points and the excitation point, and the principle of vibration response, the vibration response can be analyzed to determine the state of the elastic elements of the vibration isolators in the floating slab track bed. This allows for timely detection of problems such as steel spring breakage failure, ensuring the operational safety of the floating slab track. Furthermore, this system and method only require the installation of vibration detection devices on the upper surface of the track bed and excitation on the lower foundation; it does not require lifting the track bed slab or disassembling and assembling vibration isolators, making the detection very convenient and efficient.

[0098] Furthermore, the vibration detection device also includes a metal counterweight block, which is placed at a designated measuring point on the track bed. The vibration acceleration sensor is fixed to the counterweight block by an insulating magnetic base. This not only makes the installation and fixation of the sensor very convenient, but also enables the generation / enhancing of unbalanced vibration through the counterweight method, which is beneficial for the analysis of vibration response.

[0099] Furthermore, since each measuring point is set on the surface of the track bed slab on one side of a vibration isolator, and the line connecting the measuring point and the center of the corresponding vibration isolator is in the direction of the track width, the horizontal distance between the two is small. Therefore, the correspondence between the measuring point and the vibration isolator is intuitive and clear, and the vibration response of the measuring point can well reflect the state of the elastic element in the corresponding vibration isolator, thereby further improving the accuracy of the detection.

[0100] Furthermore, the calculation and analysis device includes a vibration signal acquisition unit, a force signal acquisition unit, a data preprocessing unit, a state parameter calculation unit, and an elastic element state determination unit. It can acquire vibration signal data from various vibration acceleration sensors and force signal data from force sensors, and preprocess the signal data to improve data quality. Then, it calculates various predetermined state parameters based on the preprocessed signal data, and automatically determines the state of the elastic element based on the state parameters and predetermined determination rules. Therefore, it has a high degree of automation. Since the calculation of state parameters and the rule-based determination are algorithms with relatively low computational load, the detection results can be obtained quickly. Thus, the detection results can be obtained on-site at the floating slab track. The detection and corresponding maintenance, replacement of elastic elements, and other operations can be completed during a single line stop maintenance, which is more efficient.

[0101] Furthermore, the calculated state parameters include the response time, energy ratio, attenuation rate, and natural frequency of the track slab at each measuring point. Based on these parameters, a primary judgment criterion and multiple auxiliary judgment criteria are established. The primary judgment criterion is set according to the response time and the expected response sequence of measuring points at different locations. The elastic element is deemed to be in an abnormal state when the primary judgment criterion is met and at least one auxiliary judgment criterion is met, or when the primary judgment criterion is not met but many auxiliary judgment criteria are met. Therefore, this comprehensive approach considers the multifaceted dynamic structural characteristics of the floating slab track, avoiding misjudgments caused by a single method and further improving detection accuracy. In addition, when the primary judgment criterion is met but all auxiliary judgment criteria are not met, the elastic element is judged to be in a pending state, further preventing the missed detection of abnormal elastic elements and ensuring safety.

[0102] Furthermore, the test metadata, raw signal data, intermediate data (preprocessed data, calculated state parameters, etc.), and judgment state information are all stored. Therefore, after multiple tests, big data related to the state detection of elastic elements can be formed, which can be used for further analysis to discover other defects or hidden dangers in the floating plate track, or as training data for machine learning models, thereby improving the level of intelligence in the industry.

[0103] <Example 2>

[0104] This embodiment provides a method and system for detecting the state of track vibration isolation elastic elements. In this embodiment, the same symbols are assigned to the same constituent elements as in Embodiment 1, and the corresponding descriptions are omitted.

[0105] Figure 8 This is a structural block diagram of the computational analysis device in this embodiment.

[0106] like Figure 8 As shown, the difference between this embodiment and Embodiment 1 is that the configuration of the computational analysis device is different in this embodiment.

[0107] The computational analysis device 140 of this embodiment includes a data storage unit 1401, a vibration signal acquisition unit 1402, a force signal acquisition unit 1403, a metadata acquisition unit 1404, a data preprocessing unit 1405, a feature extraction unit 1409, a feature fusion unit 1410, an elastic element state determination unit 1407, an analysis-side control unit 1408, an input display unit 1411, and an analysis-side communication unit 1412. The functions of the data storage unit 1401, the vibration signal acquisition unit 1402, the force signal acquisition unit 1403, the metadata acquisition unit 1404, the data preprocessing unit 1405, the input display unit 1411, and the analysis-side communication unit 1412 are basically the same as those in Embodiment 1.

[0108] The feature extraction unit 1409 is used to extract various feature parameters reflecting the structural dynamic characteristics of the floating slab track bed from the preprocessed signal data. In this embodiment, for the effective impact response signal (i.e., the preprocessed signal data), extraction is performed from multiple perspectives, including the time domain, frequency domain, and time-frequency domain. Extractable feature parameters include, but are not limited to: 1) Time domain: peak acceleration, root mean square value, crest factor, decay time, peak value and time delay of the cross-correlation function (time delay between measurement points); 2) Frequency domain: dominant frequency, peak amplitude, frequency band energy (e.g., low frequency, mid frequency, high frequency), frequency component complexity, transfer function amplitude / phase (frequency response function), coherence function (excitation and response); 3) Time-frequency domain: fast Fourier transform (FFT) energy characteristics, such as marginal spectrum energy. All of the above feature parameters can be automatically calculated, for example, using the functions provided by existing vibration and noise testing software.

[0109] The feature fusion unit 1410 is used to fuse the extracted multiple feature parameters to obtain an integrated feature vector. In this embodiment, feature parameters extracted from the time domain, frequency domain, and time-frequency domain are combined with the feature value distribution and correlation of multiple measurement points to construct a multi-dimensional feature vector. That is, the extracted multiple feature parameters and the feature value distribution and correlation coefficient of the measurement points are all put into an array as the fused feature vector.

[0110] The elastic element state determination unit 1407 is used to analyze the fused feature vector and determine the state of the elastic element based on the analysis results. For example, statistical methods (such as comparing with historical baselines or setting thresholds), pattern recognition methods (such as support vector machines (SVM) or decision trees) or machine learning models (such as machine learning models trained with historical data) can be used to analyze the fused feature vector, obtain the analysis results, and further determine the state of the elastic element.

[0111] In this embodiment, the elastic element state determination unit 1407 adopts a hybrid model of convolutional neural network and bidirectional long short-term memory network (CNN+BiLSTM). The fused features are input into the model, and the model outputs the classification result. The input data is in the format of (N, T, C), where N is the number of samples, T is the length of each time series (e.g., 10 seconds × 5000Hz = 50000 points), and C is the number of channels (i.e., the number of vibration acceleration sensors; in this embodiment, there are 3 measurement points, so C = 3). After a sample data (T, C) is input into the model, the model first extracts local temporal features through its one-dimensional convolutional layer; then the extracted features are processed by a max pooling layer; next, long-range dependencies are extracted through a bidirectional long short-term memory network (Bidirectional LSTM); then, a dropout layer is used to prevent overfitting, and then a dense fully connected layer is used; finally, the softmax layer outputs the classification results. In this embodiment, the classification results include normal state, abnormal state of elastic element corresponding to measurement point C1, abnormal state of elastic element corresponding to measurement point C2, and abnormal state of elastic element corresponding to measurement point C3.

[0112] For training the hybrid model described above, the elastic elements of the vibration isolators 230 in a track bed slab 220 can be set to a normal state, or several of them can be set to an abnormal state (e.g., a broken steel spring). Following steps S1 to S3, force signal data and three-channel vibration signal data under the corresponding states are used. The collected data is combined with the corresponding annotation information to form training samples, thus obtaining a training dataset. Preferably, for multiple elastic elements of the track bed slab 220, at least 50 samples are collected in both the normal and abnormal states. That is, key measuring points and excitation points are set next to each vibration isolator 230, and data is collected multiple times. During sample collection, the striking force of the excitation device 130 is kept as consistent as possible to control variables.

[0113] Then, the hybrid model is trained using the aforementioned training dataset. For example, PyTorch or TensorFlow can be used for training, for 100 epochs or more. During training, the classification accuracy and confusion matrix of the hybrid model are observed, and adjustments are made accordingly. After training, the hybrid model can automatically output the classification results (state determination results) for the elastic element state based on the input fused feature vector. Since the training data includes test data from key measurement points set next to each vibration isolator 230, the classification results can include: all elastic elements in a track bed slab are in a normal state; elastic element one is in an abnormal state; elastic element two is in an abnormal state, and so on.

[0114] The analysis-side control unit 1408 is used to control the operation of the above-mentioned functional units, including controlling the data storage unit 1401 to store the original signal data after the vibration signal acquisition unit 1402 and the force signal acquisition unit 1403 acquire the corresponding vibration acceleration signal data and force signal data; and sequentially controlling the data preprocessing unit 1405, the feature extraction unit 1409, the feature fusion unit 1410, and the elastic element state determination unit 1407 to perform data preprocessing, feature extraction, feature fusion, and elastic element state determination, etc.

[0115] Figure 9 This is a flowchart of the method for detecting the state of the track vibration isolation elastic element in this embodiment.

[0116] like Figure 9 As shown, based on the above-mentioned track vibration isolation elastic element state detection system 100, compared with Embodiment 1, the method in this embodiment differs only in step S4. Step S4 in this embodiment includes the following sub-steps:

[0117] Step S4-1 involves preprocessing the force signal data and the vibration signal data at each measuring point. This step is the same as step S4-1 in Example 1.

[0118] Step S4-2 involves extracting features from the preprocessed force signal data and the vibration signal data at each measuring point, extracting various feature parameters in the time domain, frequency domain, and time-frequency domain. This is the feature extraction operation performed by the feature extraction unit 1409 described above. This step is the same as step S4-2 in Embodiment 1.

[0119] Step S4-4 involves fusing the extracted feature parameters and combining the distribution and correlation information of multiple measurement points to obtain a fused feature vector. This is the feature fusion operation performed by the feature fusion unit 1410 mentioned above.

[0120] Steps S4-5 involve inputting the fused feature vectors into the trained machine learning model, which outputs the state information of the elastic element. This corresponds to the analysis and state determination performed by the elastic element state determination unit 1407, preferably using the aforementioned CNN+BiLSTM hybrid model, with the classification result serving as the state information.

[0121] In this embodiment, the other structures and method steps are the same as in Embodiment 1, so they will not be described again.

[0122] Functions and effects of Example 2

[0123] Based on the function and effect of the track vibration damping elastic element state detection method and system provided in this embodiment, since the system has a feature extraction unit and a feature fusion unit, and the elastic element state determination unit adopts a trained machine learning model, it can automatically extract a variety of feature parameters that can reflect the structural dynamic characteristics of the floating slab track bed from the signal data according to a predetermined algorithm. After fusing them, the trained machine learning model is used to realize automatic classification based on the test data to obtain the state information of the elastic element. Therefore, the efficiency of signal data analysis and processing and state determination is higher, thereby achieving more efficient detection and further reducing downtime for maintenance.

[0124] Furthermore, the machine learning model preferably employs a CNN + BiLSTM hybrid model. This model takes a multi-dimensional array as input, thus supporting input of multi-channel vibration signal data (multi-sensor data), making it more convenient. Moreover, this model has excellent sequence dependency capture capabilities; since abnormal states of elastic elements (broken springs) affect the vibration transmission sequence, this model is highly suitable for analysis and judgment in such scenarios. In addition, this model achieves strong classification accuracy and robustness without complex manual feature engineering.

[0125] Furthermore, since force and vibration signal data are collected in the actual floating slab track as training data, rather than using simulated data, the trained model has high adaptability and classification accuracy for determining the state of elastic elements in the actual track.

[0126] The above embodiments are merely illustrative of specific implementations of the present invention, and the present invention is not limited to the scope of the description of the above embodiments. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are only for illustrating the principles of the present invention. Various changes and modifications can be made to the present invention without departing from the spirit and scope thereof, and all such changes and modifications fall within the scope of the present invention as claimed. The scope of protection of the present invention is defined by the appended claims and their equivalents.

[0127] For example, in the above embodiments, the method and system for detecting the state of track vibration damping elastic elements are used to detect the state of steel springs, especially to detect the spring breakage of steel springs. In alternative embodiments, the method and system can also be used similarly to detect the state of other types of track vibration damping elastic elements, such as rubber springs and combined springs.

Claims

1. A track vibration isolation elastic element condition detection system, installed in a floating slab track having a subbase, track bed slab, and multiple vibration isolators, for detecting the condition of the elastic elements in the vibration isolators, characterized in that, include: Multiple vibration detection devices are set at predetermined measuring points in the floating slab track to acquire vibration signal data; The excitation device is used to excite predetermined excitation points in the floating plate track and acquire corresponding force signal data; as well as The computational analysis device is used to analyze the vibration response to excitation based on force signal data and vibration signal data from various measuring points, and to determine the state of the elastic element based on the analysis results. The measuring points include at least one key measuring point and one adjacent measuring point. The key measuring point is located on the upper surface of the track slab on one side of one of the vibration isolators, and the adjacent measuring point is located on the upper surface of the track slab on the other side of the vibration isolator adjacent to that vibration isolator. The excitation point is located on the lower foundation of the vibration isolator side corresponding to the key measuring point. The state parameter calculation unit in the computational analysis device calculates various state parameters based on the preprocessed vibration signal data and force signal data. The elastic element state determination unit in the calculation and analysis device determines the state of the elastic element based on state parameters and predetermined determination rules. The state parameters include at least the response time of each measuring point to the excitation, the energy ratio of the force signal data and the vibration signal data, the attenuation rate of the vibration signal, and the natural frequency of the track slab including the vibration isolator. The judgment rules include: when the main judgment criteria are met and at least one auxiliary judgment criterion is met, or when the main judgment criteria are not met but more than a predetermined number of auxiliary judgment criteria are met, the elastic element corresponding to the key measuring point is judged to be in an abnormal state. The main criterion is that the response time of the key measuring point and the response time of adjacent measuring points do not meet the predetermined response order. Multiple auxiliary judgment criteria include: the energy relative ratio of key measuring points and the energy relative ratio of adjacent measuring points do not conform to the predetermined energy ratio relationship; the attenuation rate deviates from the predetermined attenuation rate benchmark value by more than a predetermined threshold; the natural frequency deviates from the predetermined natural frequency benchmark value by more than a predetermined threshold. The response order is as follows: the response time of all nearby measuring points is greater than the response time of the critical measuring point. The energy ratio relationship is as follows: the relative energy ratio of all nearby measuring points is less than the relative energy ratio of the key measuring point.

2. The track vibration isolation elastic element condition detection system according to claim 1, characterized in that, Also includes: Signal acquisition device The vibration detection device includes a vibration acceleration sensor. The excitation device includes a force sensor. The signal acquisition device is connected to the vibration acceleration sensor, the force sensor, and the calculation and analysis device, respectively, and is used to acquire the vibration signal data and the force signal data and transmit them to the calculation and analysis device. The vibration detection device further includes: A metal counterweight is positioned at the measuring point; and An insulating magnetic base is attached to the counterweight, and the vibration acceleration sensor is fixed to the counterweight via the insulating magnetic base. The excitation device is a force hammer with the force sensor built in.

3. The track vibration isolation elastic element condition detection system according to claim 2, characterized in that: in, The number of nearby measuring points is one or two. The measuring point is located on the side of the corresponding vibration isolator near the middle of the track slab. The horizontal direction connecting the measuring point to the center of the corresponding vibration isolator, and the horizontal direction connecting the key measuring point to the excitation point, are both along the width direction of the floating slab track. The horizontal distance between the measuring point and the center of the corresponding vibration isolator is 15cm to 25cm. The lower foundation is the tunnel wall, and the vertical distance between the excitation point and the key measuring point is 1m to 1.8m.

4. The track vibration isolation elastic element condition detection system according to claim 2, Its features are: The computational analysis device further includes: The vibration signal acquisition unit is used to acquire corresponding vibration signal data from each of the vibration acceleration sensors; Force signal acquisition unit, used to acquire force signal data from the force sensor; and The data preprocessing unit is used to preprocess the vibration signal data and the force signal data.

5. The track vibration isolation elastic element condition detection system according to claim 2, Its features are: The computational analysis device further includes: The vibration signal acquisition unit is used to acquire corresponding vibration signal data from each of the vibration acceleration sensors; A force signal acquisition unit is used to acquire force signal data from the force sensor; A data preprocessing unit is used to preprocess the vibration signal data and the force signal data; The feature extraction unit is used to extract features from the preprocessed vibration signal data and force signal data to obtain various feature parameters that reflect the structural dynamic characteristics of the floating slab track bed; and The feature fusion unit is used to fuse multiple feature parameters and combine the distribution and correlation information of multiple measurement points to obtain a fused feature vector. The elastic element state determination unit further processes the fused feature vector using a trained machine learning model and determines the state of the elastic element.

6. The track vibration isolation elastic element condition detection system according to claim 5, characterized in that: in, The feature extraction unit extracts feature parameters in the time domain, frequency domain, and time-frequency domain. The time-domain characteristic parameters include peak acceleration, root mean square value, crest factor, decay time, peak value of cross-correlation function, and time delay. The characteristic parameters in the frequency domain include the dominant frequency, peak amplitude, bandwidth energy, frequency component complexity, transfer function amplitude / phase, and coherence function. The time-frequency domain characteristic parameters include the Fast Fourier Transform energy characteristics. The feature fusion unit integrates the feature parameters in the time domain, frequency domain, and time-frequency domain, and combines them with the distribution and correlation information of the measurement points to obtain a fused feature vector. The machine learning model is a hybrid model of convolutional neural network and bidirectional long short-term memory network, which is used to output the classification result of the state of the elastic element based on the input fused feature vector.

7. The track vibration isolation elastic element condition detection system according to claim 6, characterized in that: in, Multiple vibration isolators corresponding to one track slab are set to normal state, and several of the multiple vibration isolators corresponding to one track slab are set to abnormal state. Key measuring points and excitation points are set on one side of each vibration isolator in sequence. Vibration signal data and force signal data are collected respectively and corresponding label information is set to be used as training data to train the machine learning model.

8. A method for detecting the state of a track vibration isolation elastic element, characterized in that, include: Step S1: Install a track vibration isolation elastic element condition detection system as described in any one of claims 1-7 in a floating slab track having a lower foundation, track bed slab and multiple vibration isolators, wherein the vibration isolators include elastic elements; Step S2: Use the excitation device to excite at the excitation point and acquire the force signal data during excitation; Step S3: Vibration signal data at each of the measuring points are collected by the multiple vibration detection devices. Step S4: Analyze the vibration response to the excitation based on the force signal data and the vibration signal data of each of the measuring points, and determine the state of the elastic element based on the analysis results.

Images