Track fastener high-frequency dynamic stiffness identification method, program product, electronic equipment and storage medium

By obtaining vertical vibration acceleration data of the axle box and using the unsprung mass-rail system analysis model, the high-frequency dynamic stiffness of the track fastener is solved in real time, and the problem of difficulty in monitoring the track vibration level in the existing technology is realized, and railway operation is achieved safety and stability.

CN118897980APending Publication Date: 2024-11-05SOUTHWEST JIAOTONG UNIV

Patent Information

Application Number
CN202411160592.3
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-08-22
Publication Date
2024-11-05

AI Technical Summary

Technical Problem

The prior art lacks a method that can identify the high-frequency dynamic stiffness of rail fasteners in real time, efficiently and stably, making it difficult to monitor and control the track vibration level when the train is driving, affecting the maintenance and operation of the track system.

Method used

By obtaining the vertical vibration acceleration data of the axle box and using the unsprung mass-rail system analysis model, the transcendent equation is solved to determine the first-order natural frequency of the wheel rail, thereby identifying the high-frequency dynamic stiffness of the track fastener based on the Winkler foundation model. , empirical modal decomposition method and N4SID subspace recognition method are used to process data to achieve real-time recognition.

Benefits of technology

This method does not require additional testing equipment, and can identify the high-frequency dynamic stiffness of fasteners in real service in real time and accurately, ensure the safety and long-term stability of railway operations, and improve the efficiency and accuracy of vibration level monitoring of the track system.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118897980A_ABST
    Figure CN118897980A_ABST
Patent Text Reader

Abstract

The invention provides a track fastener high-frequency dynamic stiffness identification method, a program product, electronic equipment and a storage medium, and the method comprises the steps: obtaining vertical vibration acceleration data of a train axle box on a target track along a time sequence, and obtaining an unsprung mass-track system analysis model, the unsprung mass-track system analysis model is used for analyzing the dynamic performance of a target track based on a Winkler foundation model, and comprises a transcendental equation taking the supporting stiffness of a steel rail per unit length as an unknown parameter; determining a first-order inherent frequency of a wheel track corresponding to the target track according to the vertical vibration acceleration data; and inputting the first-order inherent frequency into the unsprung mass-track system analysis model, solving the transcendental equation to obtain the bearing stiffness, and identifying the high-frequency dynamic stiffness of the track fastener corresponding to the target track according to the bearing stiffness. The method can identify the high-frequency dynamic stiffness of the fastener on the track site in real time, does not need additional equipment, does not damage the track structure, and is wide in monitoring range.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present application relates to the field of rail transit technology, and in particular to a method for identifying high-frequency dynamic stiffness of rail fasteners, a program product, an electronic device and a storage medium. Background Art

[0002] During the track operation stage, the elastic elements in the fastener system may deteriorate due to factors such as long-term cyclic loads of trains, hot and humid environments, and improper maintenance. This degradation often leads to an increase in the high-frequency dynamic stiffness of the fasteners, which in turn significantly increases the vibration level of the track when the train is running. For example, some tracks did not cause complaints from residents due to vibration problems in the early stages, but as time goes by, the vibration level may exceed the standard, causing a large number of complaints and protests from residents. Therefore, during the operation and maintenance of the track, monitoring and controlling the high-frequency dynamic stiffness of the fasteners becomes a key link in ensuring the vibration level of the track system, maintaining track performance, and reducing maintenance costs.

[0003] Currently, there is a lack of a method that can identify the dynamic stiffness of rail fasteners in real time, efficiently and stably. Summary of the invention

[0004] The purpose of the embodiments of the present application is to provide a method for identifying the high-frequency dynamic stiffness of a rail fastener, a program product, an electronic device and a storage medium, so as to identify the high-frequency dynamic stiffness of the track on site in real time.

[0005] In a first aspect, the present application provides a method for identifying the high-frequency dynamic stiffness of a track fastener, comprising: obtaining vertical vibration acceleration data of a train axle box on a target track along a time series, and obtaining an unsprung mass-track system analysis model, wherein the unsprung mass-track system analysis model is used to analyze the dynamic performance of the target track based on a Winkler foundation model, including a transcendental equation with the support stiffness of a unit length of rail as an unknown parameter; determining a first-order natural frequency of a wheel-rail corresponding to the target track according to the vertical vibration acceleration data; inputting the first-order natural frequency into the unsprung mass-track system analysis model, solving the transcendental equation to obtain the support stiffness, and identifying the high-frequency dynamic stiffness of the track fastener corresponding to the target track according to the support stiffness.

[0006] In the above scheme, the vertical vibration acceleration data collected by the acceleration sensors installed in the train axle boxes are creatively introduced into the field of high-frequency dynamic stiffness identification. The vertical vibration acceleration data in the train axle boxes are collected in real time, so the identification of the high-frequency dynamic stiffness of the rail fasteners is also carried out in real time, and the high-frequency dynamic stiffness of the fasteners in the actual service state can be measured. In addition, this identification method does not require the addition of additional testing equipment and has no destructive effect on the track structure, thereby ensuring the safety and long-term stability of railway operations.

[0007] As an optional way, before obtaining the unsprung mass-track system analysis model, the method further includes: constructing a vibration differential equation of the target track, the vibration differential equation characterizing the dynamic performance of the rail and the wheel during vibration when the train wheel load acts on any point on the target track; obtaining the vibration mode function of the rail according to the vibration differential equation; and obtaining the transcendental equation according to the natural frequency characteristic equation of the vibration mode function. In the above scheme, on the basis of representing the target track with the Winkler foundation model, the vibration mode function of the rail is obtained by establishing the vibration differential equation of the system, and finally the transcendental equation of stiffness is obtained.

[0008] As an optional method, the expression of the transcendental equation is: Among them, M U is half of the wheelset mass, k H is the linearized wheel-rail contact stiffness, EI is the bending stiffness of the rail, k is the support stiffness, m r is the mass of the rail per unit length, L is the length of the rail, Among them, α 2 In the above scheme, the first-order natural frequency and other on-site track parameters that can be obtained in advance are substituted into the transcendental equation, and the high-frequency dynamic stiffness at the first-order natural frequency of the wheel-rail can be calculated.

[0009] As an optional method, the first-order natural frequency of the wheel-rail corresponding to the target track is determined according to the vertical vibration acceleration data, including: using the empirical mode decomposition method to extract all the natural mode functions from the vertical vibration acceleration data; using the N4SID subspace identification method to determine the state space model corresponding to the natural mode function, and determining the first-order natural frequency from the state space model. In the above scheme, the empirical mode decomposition method and the N4SID subspace identification method are used to process and analyze the vertical vibration acceleration data to finally obtain the first-order natural frequency of the wheel-rail, laying the foundation for the subsequent identification of high-frequency dynamic stiffness.

[0010] As an optional manner, the empirical mode decomposition method is used to extract all intrinsic modal functions from the vertical vibration acceleration data, including: identifying local maximum values ​​and local minimum values ​​in the vertical vibration acceleration data, and constructing an envelope between the local maximum values ​​and local minimum values; extracting the first intrinsic modal function from the vertical vibration acceleration data according to the envelope, and subtracting the first intrinsic modal function from the vertical vibration acceleration data to obtain remaining data; repeating the following steps until the remaining data contains no more than two extreme values: extracting a new intrinsic modal function from the remaining data, and subtracting the new intrinsic modal function from the remaining data to obtain new remaining data.

[0011] In the above scheme, the empirical mode decomposition method is used to extract the intrinsic mode function, which can adaptively process nonlinear and non-stationary vertical vibration acceleration data without relying on preset basis functions, and can effectively analyze the natural frequency of the wheel-rail.

[0012] As an optional manner, the N4SID subspace identification method is used to determine the state space model corresponding to the intrinsic mode function, and determine the first-order natural frequency from the state space model, including: constructing a past output matrix and a future output matrix based on the intrinsic mode function, the past output matrix represents the output data of the system in a period of time in the past, and the future output matrix represents the output data of the hypothetical system in a period of time in the future; based on the past output matrix, using LQ decomposition to calculate the spatial projection of the future output matrix; according to the spatial projection, using singular value decomposition method to calculate the future state vector; according to the future state vector, determining the state space model, and using the least squares method to determine the dynamic matrix of the state space model; determining the first-order natural frequency according to the dynamic matrix.

[0013] In the above scheme, the N4SID subspace identification technology is used to analyze the intrinsic mode function to determine the natural frequency corresponding to the model determined by the intrinsic mode function, including the first-order natural frequency of the wheel-rail.

[0014] As an optional method, after obtaining the vertical vibration acceleration data of the train axle box on the target track along the time series, the method further includes: filtering and normalizing the vertical vibration acceleration data. In the above scheme, in order to ensure that the vertical vibration acceleration data along the time series is smoother and more stable and improve the data quality, the data is filtered and normalized.

[0015] In a second aspect, the present application further provides a computer program product, comprising computer program instructions, wherein when the computer program instructions are read and executed by a processor, the method according to the first aspect is executed.

[0016] In a third aspect, the present application provides an electronic device comprising: a processor, a memory and a bus, wherein the processor and the memory communicate with each other through the bus; the memory stores program instructions executable by the processor, and the processor calls the program instructions to execute the method described in the first aspect above.

[0017] In a fourth aspect, the present application provides a computer-readable storage medium, comprising: the computer-readable storage medium stores computer instructions, and the computer instructions enable the computer to execute the method described in the first aspect above.

[0018] In a fifth aspect, the present application provides a device for identifying the high-frequency dynamic stiffness of a track fastener, comprising: an acquisition module for acquiring vertical vibration acceleration data of a train axle box on a target track along a time series, and acquiring an unsprung mass-track system analysis model, wherein the unsprung mass-track system analysis model is used to analyze the dynamic performance of the target track based on a Winkler foundation model, including a transcendental equation with the support stiffness of a unit length of rail as an unknown parameter; a first-order natural frequency determination module for determining the first-order natural frequency of a wheel-rail corresponding to the target track according to the vertical vibration acceleration data; and a high-frequency dynamic stiffness identification module for inputting the first-order natural frequency into the unsprung mass-track system analysis model, solving the transcendental equation to obtain the support stiffness, and identifying the high-frequency dynamic stiffness of the track fastener corresponding to the target track according to the support stiffness.

[0019] Other features and advantages of the present application will be described in the subsequent description, and in part will become apparent from the description, or will be understood by implementing the embodiments of the present application. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings required for use in the embodiments of the present application will be briefly introduced below. It should be understood that the following drawings only show certain embodiments of the present application and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without paying creative work.

[0021] Figure 1 A schematic flow chart of a method for identifying high-frequency dynamic stiffness of a rail fastener provided in an embodiment of the present application;

[0022] Figure 2 A schematic diagram of a Winkler foundation beam-wheel system model in an embodiment of the present application;

[0023] Figure 3A design flow chart of a method for identifying high-frequency dynamic stiffness of a rail fastener provided in an embodiment of the present application;

[0024] Figure 4 A set of vertical vibration acceleration data of axle boxes of trains tested at around 8 o'clock in the morning in a certain city's subway provided in the embodiment of the present application;

[0025] Figure 5 A set of axle box vertical vibration acceleration data before filtering and normalization provided in the embodiment of the present application;

[0026] Figure 6 A set of axle box vertical vibration acceleration data after filtering and normalization provided in the embodiment of the present application;

[0027] Figure 7 The wheel-rail first-order natural mode function extracted from a set of axle box vertical vibration acceleration data using the empirical mode decomposition method provided in the embodiment of the present application;

[0028] Figure 8 The first-order natural frequency of the wheel / rail identified from 20 sets of axle box vertical vibration acceleration data using the method provided in the embodiment of the present application is shown;

[0029] Fig. 9 The figure shows the high frequency dynamic stiffness of fasteners identified from 20 sets of axle box vertical vibration acceleration data using the method provided in the embodiment of the present application;

[0030] Fig.10 A schematic diagram showing a comparison between the high-frequency dynamic stiffness of fasteners identified from 20 sets of axle box vertical vibration acceleration data using the method provided in an embodiment of the present application and the indoor test reference value;

[0031] Fig.11 A schematic diagram of the structure of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0032] The following embodiments of the technical solution of the present application will be described in detail in conjunction with the accompanying drawings. The following embodiments are only used to more clearly illustrate the technical solution of the present application, and are therefore only used as examples, and cannot be used to limit the scope of protection of the present application.

[0033] It should be noted that all technical and scientific terms used in this document have the same meanings as those generally understood by technicians in the technical field to which this application belongs; the terms used in this document are only for the purpose of describing specific embodiments and are not intended to limit this application; the terms "including" and "having" in the specification and claims of this application and the above-mentioned figure descriptions and any variations thereof are intended to cover non-exclusive inclusions.

[0034] In the description of the embodiments of the present application, “plurality” means more than two, unless otherwise clearly and specifically defined.

[0035] In the description of the embodiments of the present application, the term "and / or" is only a description of the association relationship of the associated objects, indicating that there may be three relationships. For example, A and / or B can represent: A exists alone, A and B exist at the same time, and B exists alone. In addition, the character " / " in this article generally indicates that the associated objects before and after are in an "or" relationship.

[0036] The basic principle of the existing track dynamic stiffness detection scheme is to connect the exciter and the rail through a fixing device, and apply dynamic excitation to the rail through the exciter. The magnitude of the dynamic force can be read in real time by the force sensor connected to the exciter, and the displacement of the rail can be directly obtained by the displacement sensor connected to the rail. The dynamic secant stiffness of the track can be obtained based on the force and displacement. However, this method requires the addition of test equipment such as the exciter to be connected to the rail, and maintenance personnel are required to conduct on-site testing, which increases the workload of maintenance and repair. In addition, this test is the high-frequency dynamic stiffness of the fastener obtained when there is no train, and it is impossible to measure the actual service stiffness of the fastener when the train passes.

[0037] After discovering the above problems in the prior art, the inventors proposed a method for identifying the high-frequency dynamic stiffness of a rail fastener, in which no additional testing equipment is required. The following is a detailed introduction to the method for identifying the high-frequency dynamic stiffness of a rail fastener provided in an embodiment of the present application.

[0038] Please refer to Figure 1 , Figure 1 A schematic flow chart of a method for identifying high-frequency dynamic stiffness of a rail fastener provided in an embodiment of the present application, the method comprising the following steps:

[0039] Step S10: obtaining vertical vibration acceleration data of axle boxes of trains on the target track along a time series, and obtaining an unsprung mass-track system analysis model, wherein the unsprung mass-track system analysis model is used to analyze the dynamic performance of the target track based on the Winkler foundation model, including a transcendental equation with the support stiffness of the rail per unit length as an unknown parameter.

[0040] In railway operations, when each line is opened, acceleration sensors are installed in the train axle boxes to monitor the train operation status. The vertical vibration acceleration data used in the embodiment method of the present application comes from the train axle boxes. This is also the reason why the high-frequency dynamic stiffness identification method provided in the present application does not require additional testing equipment.

[0041] Unsprung mass refers to the mass of the part of the rail vehicle that is not supported by the suspension system, that is, the mass of the part that is in direct contact with the track, mainly including wheels, axles, and some bogie components. The unsprung mass-track system analysis model considers the target track system as a homogeneous rail with a length of L on the Winkler foundation. Based on this, a Winkler foundation beam-wheel system model is established to analyze the vibration characteristics of the target track when interacting with the train wheel. Since the parameter concerned by the unsprung mass-track system analysis model is stiffness, a transcendental equation that regards the support stiffness of the unit length of the rail as an unknown parameter can be derived from the vibration characteristics.

[0042] It should be noted that the vertical vibration acceleration data can be data collected in real time on site, or data in a directly accessible state, such as vertical vibration acceleration data received from an external device, read from the cloud or stored locally in advance, etc. The unsprung mass-track system analysis model can be a model established based on parameters measured on site, or data in a directly accessible state, such as vertical vibration acceleration data received from an external device, read from the cloud or stored locally in advance, etc.

[0043] Step S20: determining the first-order natural frequency of the wheel-rail corresponding to the target track according to the vertical vibration acceleration data.

[0044] The method provided in this application only relies on the vertical vibration acceleration data collected in real time by the existing acceleration sensor on the train axle box. By performing spectrum analysis on these acceleration data, the first-order natural frequency of the wheel-rail can be quickly and accurately identified. As another implementation method, more complex data analysis techniques such as Empirical Mode Decomposition (EMD) and N4SID (Numerical Algorithms for Subspace State Space System Identification) subspace identification methods can also be used to extract the first-order natural frequency of the wheel-rail.

[0045] Step S30: inputting the first-order natural frequency into the unsprung mass-track system analysis model, solving the transcendental equation to obtain the support stiffness, and identifying the high-frequency dynamic stiffness of the track fastener corresponding to the target track according to the support stiffness.

[0046] The parameters in the transcendental equation include the first-order natural frequency. Therefore, after substituting the first-order natural frequency obtained in step S20, the rail support stiffness per unit length corresponding to the first-order natural frequency of the wheel-rail can be solved. Considering the distribution spacing of the actual rail fasteners, the high-frequency dynamic stiffness of the rail fasteners at the first-order natural frequency of the wheel-rail can be obtained.

[0047] In the above scheme, the vertical vibration acceleration data collected by the acceleration sensors installed in the train axle boxes are creatively introduced into the field of high-frequency dynamic stiffness identification. The vertical vibration acceleration data in the train axle boxes are collected in real time, so the identification of the high-frequency dynamic stiffness of the rail fasteners is also carried out in real time, and the high-frequency dynamic stiffness of the fasteners in the actual service state can be measured. In addition, this identification method does not require the addition of additional testing equipment and has no destructive effect on the track structure, thereby ensuring the safety and long-term stability of railway operations.

[0048] In some embodiments, before obtaining the unsprung mass-track system analysis model, the method for identifying the high-frequency dynamic stiffness of the track fastener provided in the embodiment of the present application further includes:

[0049] Step 1): Construct a vibration differential equation of the target track, which characterizes the dynamic performance of the rail and the wheel during vibration when the train wheel load acts on any point on the target track.

[0050] like Figure 2 As shown, Figure 2 Schematic diagram of the Winkler foundation beam-wheel system model in the embodiment of the present application. A two-dimensional plane coordinate system is established with the starting point of the rail as the origin, and the extension direction of the rail is the x-axis. The Winkler foundation model essentially regards the foundation as many unconnected vertically arranged springs. o It includes the primary spring force of the train vehicle and the gravity caused by the wheel mass, k H is the linearized wheel-rail contact stiffness, k is the support stiffness per unit length of rail, M U is half of the mass of the wheelset, that is, half of the mass of an axle and the two wheels at both ends, L is the length of the rail, when the wheel load acts at x = 0, the vibration differential equation of the system can be expressed as formula (1) and formula (2):

[0051]

[0052] M U y U ”+k H [y U -y(x0,t)]=P0 (2)

[0053] In the above formula, EI is the bending stiffness of the rail, m r The mass of the rail per unit length, k is the support stiffness of the rail per unit length, M U is half of the wheelset mass, k H is the linearized wheel-rail contact stiffness, y and y U represents the vertical displacement of the rail and the wheel, P o It includes the primary spring force of the vehicle and the gravity caused by the wheel mass. δ is the Dirac function. PU is the wheel-rail vertical force, expressed as formula (3):

[0054] P U =k H [y U -y(x0,t)]=P0-M U y U ” (3)

[0055] Step 2): According to the vibration differential equation, the vibration mode function of the rail is obtained.

[0056] Considering the problem of free vibration of the system, we can let P o =0, Combining formulas (1) to (3) and solving them using the separation of variables method, we can obtain formula (4):

[0057]

[0058] In formula (4), Y(x) is the vibration mode function of the rail to be determined, α 2 is the natural frequency to be determined, α 4 =(λ 4 EI+k) / m r ,

[0059] By performing the inverse Laplace transform on formula (4), we can get formula (5). In formula (5), G(x) is the bending response curve of the rail under the unit impulse force. G(x) is expressed by formula (6), where u(x-0) is the unit step function. The coefficients A, B, C, and D can be determined by the boundary conditions of the rail beam model:

[0060]

[0061] G(x)=Asinλx+Bcosλx+Csinhλx+Dcoshλx+

[0062] (1 / 2λ 3 )[sinhλ(x-x0)-sinλ(x-x0)]u(x-x0)(6)

[0063] When the track is considered as simply supported at both ends, the vibration mode function of the rail can be obtained as formula (7):

[0064]

[0065] Step 3): According to the natural frequency characteristic equation of the vibration mode function, the transcendental equation is obtained.

[0066] When x=0, Y(x)=Y(x0), substituting into (7) yields the system's natural frequency characteristic equation (8):

[0067]

[0068] From this, a series of eigenvalues ​​λ of the system can be obtained i , i=1, 2, ..., the first-order natural frequency corresponding to λ1 is the first-order natural frequency of the wheel-rail.

[0069] Since the method of the embodiment of the present application focuses on the first-order natural frequency of the wheel / rail, that is, the natural frequency α to be determined 2 is the first-order natural frequency, at this time α 2 must be less than the natural frequency of the track Then from formula (7), we can know that the eigenvalue λ i The value form of is formula (9):

[0070]

[0071] Taking x0 = L / 2, equation (8) can be simplified to obtain a transcendental equation.

[0072] In the above scheme, on the basis of representing the target track with the Winkler foundation model, the vibration differential equation of the system is established to obtain the vibration mode function of the rail, and finally the transcendental equation of stiffness is obtained.

[0073] In some embodiments, the expression of the transcendental equation is:

[0074]

[0075] Among them, M U is half of the wheelset mass, k H is the linearized wheel-rail contact stiffness, EI is the bending stiffness of the rail, k is the support stiffness, m r is the mass of the rail per unit length, L is the length of the rail, α 4 =(λ 4 EI+k) / m r , where α 2 is the first-order natural frequency.

[0076] In the transcendental equation, the support stiffness k is the quantity to be determined. In addition to k, the first-order natural frequency α 2 and other parameters need to be input into the transcendental equation to solve for k.

[0077] In the above scheme, the first-order natural frequency and other on-site track parameters that can be obtained in advance are substituted into the transcendental equation, and finally the high-frequency dynamic stiffness at the first-order natural frequency of the wheel-rail can be calculated.

[0078] In some embodiments, step S20 includes:

[0079] S201: Using the empirical mode decomposition method, all inherent mode functions are extracted from the vertical vibration acceleration data.

[0080] EMD is an advanced data analysis technique designed specifically for processing nonlinear and non-stationary time series data. It effectively addresses the limitations of traditional linear analysis methods in processing such data by extracting intrinsic mode functions (IMFs) from complex data sets. The adaptability of EMD lies in its ability to extract intrinsic mode functions directly from the data itself without relying on preset basis functions, making it particularly powerful in analyzing complex signals.

[0081] Explanatory, the brief steps of extracting IMFs by EMD are: identify extreme points, find all local maxima and local minima in the original signal; fit the envelope, use the extreme points to construct the upper and lower envelopes of the original signal; calculate the local mean, calculate the average of the upper and lower envelopes, and obtain the local mean function; extract IMF, subtract the local mean function from the original signal to obtain the first IMF, then subtract the first IMF from the original signal as the new original signal, and then go through the above steps to obtain the second IMF. This process is iterated until the new original signal becomes a residual or monotonic function approaching zero.

[0082] S202: Using the N4SID subspace identification method, determine the state space model corresponding to the intrinsic mode function, and determine the first-order intrinsic frequency from the state space model.

[0083] The N4SID subspace identification method is a method for system identification. It is based on the geometric properties of linear systems and estimates the state space model of the system by processing the input and output data of the system.

[0084] Explanatory, the brief steps of the N4SID subspace identification method are: Collect input and output data: First, it is necessary to collect the time series data of the system's input and output; Construct the Hankel matrix, and construct the past input Hankel matrix U according to the system's input and output data. p 、The past output Hankel matrix Y p 、Future input Hankel matrix U f and the future output Hankel matrix Y f, these matrices contain information about the system's historical data and future forecast data; projection and singular value decomposition (SVD), weighted projection of the Hankel matrix, and SVD decomposition of the weighted projection matrix to extract the key features of the system. The result of SVD decomposition is used to estimate the state sequence or observation matrix of the system; solve the system matrix, use the result of SVD decomposition and the Hankel matrix to construct the system's Toeplitz matrix, use the Toeplitz matrix and the system's input and output data, and solve the system's state space model parameters, that is, the system matrices A, B, C and D, through least squares and other methods.

[0085] After obtaining the state-space model of the system, the first-order natural frequency is determined from the state-space model through dynamic eigenvalue analysis.

[0086] In the above scheme, the empirical mode decomposition method and N4SID subspace identification method are used to process and analyze the vertical vibration acceleration data, so as to finally obtain the first-order natural frequency of the wheel / rail, laying the foundation for the subsequent identification of high-frequency dynamic stiffness.

[0087] In some embodiments, step S201 includes:

[0088] Step 1): Identify local maximum and local minimum in vertical vibration acceleration data, and construct an envelope between the local maximum and local minimum.

[0089] After extracting the vertical vibration acceleration data, the vertical vibration acceleration data is used as the original data to identify all local maxima and minima in the original data. Then, the upper envelope and lower envelope of the original data are fitted between the maximum and minimum points respectively using the interpolation method.

[0090] Step 2): extracting the first intrinsic mode function from the vertical vibration acceleration data according to the envelope, and subtracting the first intrinsic mode function from the vertical vibration acceleration data to obtain the remaining data.

[0091] Next, the upper and lower envelopes are averaged to obtain the mean envelope, which is then subtracted from the original data to generate the first candidate for the intrinsic mode function. If the candidate meets the conditions of the intrinsic mode function (IMF), it is confirmed as an IMF, otherwise it is regarded as new original data and the envelope is refitted until the first IMF that meets the conditions is obtained.

[0092] There are two conditions for judging whether a candidate meets the IMF: first, the number of local extreme points and the number of zero-crossing points must be equal or differ by at most one; second, at any time, the average value of the upper envelope formed by the local maximum points and the lower envelope formed by the local minimum points is zero, that is, the upper and lower envelopes are locally symmetric with respect to the time axis.

[0093] When the first IMF that meets the conditions is screened out, the original data is subtracted from the first IMF to obtain the remaining data.

[0094] Step 3): Repeat the following steps until the remaining data contains no more than two extreme values: extract a new intrinsic mode function from the remaining data, and subtract the new intrinsic mode function from the remaining data to obtain new remaining data.

[0095] The extracted IMFs are continuously subtracted from the initial raw data, and the above process of identifying local extreme points, fitting envelopes, and extracting IMFs is repeated for the remaining data until the remaining data contains no more than two extreme values.

[0096] In summary, the empirical mode decomposition method not only reveals the instantaneous dynamic behavior of time series signals, but also helps analyze and predict the performance changes of the system under different conditions through the decomposition of a series of intrinsic mode functions, providing an effective analytical method for understanding and dealing with the natural frequency of the system.

[0097] In the above scheme, the empirical mode decomposition method is used to extract the intrinsic mode function, which can adaptively process nonlinear and non-stationary vertical vibration acceleration data without relying on preset basis functions, and can effectively analyze the natural frequency of the wheel-rail.

[0098] In some embodiments, step S202 includes:

[0099] Step 1): construct a past output matrix and a future output matrix based on the intrinsic mode function, wherein the past output matrix represents the output data of the system in the past period of time, and the future output matrix represents the output data of the assumed system in the future period of time.

[0100] For the intrinsic mode functions (IMFs) extracted from the empirical mode decomposition, the N4SID algorithm constructs the intrinsic mode function (IMFs) data as the past output matrix Y p and the future output matrix Y f , where Y p =Y(IMFs) 0~k-1 , Y f =Y(IMFs) k~2k-1 Since the excitation generated by the train is unknown, the past input matrix U in the identification procedure p With the future input matrix U fSet to 0, W p The matrix is ​​composed of the past input matrix and the past output matrix, defined as W p =[U p Y p ] T .

[0101] Step 2): Based on the past output matrix, LQ decomposition is used to calculate the spatial projection of the future output matrix.

[0102] Calculate Y using the LQ decomposition in formula (11) and formula (12) f To W p Along U f The purpose of calculating the spatial projection γ is to extract information related to the system dynamics from the output data. 32 With M 22θ is obtained from the LQ decomposition, where M 22θ It is M 32 The pseudo-inverse of Q k is the extended observable matrix, X f is the future state vector.

[0103] γ=M 32 M 22θ W p =Q k X f (11)

[0104]

[0105] Formula (12) is the expression of LQ decomposition, which is used to decompose the data matrix into the product of two matrices L and Q, where L is a lower triangular matrix and Q is an orthogonal matrix. LQ decomposition helps to separate the dynamic modes of the system in the subspace identification method.

[0106] Step 3): According to the spatial projection, the singular value decomposition method is used to calculate the future state vector.

[0107] The following formulas (13), (14), and (15) are used to calculate the future state vector X by SVD: f , the SVD method decomposes the spatial projection γ into U1, ∑1 and Three matrices, through which the extended observable matrix Q is calculated k and the future state vector X f , where T is a transformation matrix. Thus, X f It contains the state information of the system from the current moment to multiple moments in the future, providing a basis for predicting and controlling the behavior of the system.

[0108]

[0109] Step 4): Determine the state space model according to the future state vector, and determine the dynamic matrix of the state space model using the least squares method.

[0110] The expression of the state space model is expressed by the following formula (16). The least squares method is used to calculate the matrices A, B, C and D in the state space model. These matrices can map the current state and input of the system to the future state and output. Among them, the A matrix is ​​the dynamic matrix of the system, which describes how the state evolves over time, the B matrix is ​​the input matrix, which indicates how the input affects the evolution of the state, the C matrix is ​​the output matrix, which describes how the state produces the output, and the D matrix is ​​the direct matrix, which indicates how the input directly affects the output. In formula (16), Represents the current state, represents the predicted future state, is the future output predicted based on current information, is the future input given the current information. The following matrix is ​​constructed by combining the current state estimate, future state estimate, and future input, and the least squares method is applied to estimate the above four matrices A, B, C, and D.

[0111]

[0112] Step 5): Determine the first-order natural frequency based on the dynamic matrix.

[0113] For the intrinsic mode functions (IMFs) decomposed by the empirical mode decomposition method, the two-dimensional discrete time state space model can be defined as the following formula (17) and formula (18):

[0114]

[0115] in, Represents future output.

[0116] The A matrix is ​​a dynamic matrix in discrete time, according to the natural frequency w n The dynamic matrix defined by the damping coefficient ε is as follows:

[0117]

[0118] Therefore, according to formula (19), through dynamic eigenvalue analysis, the resonance frequency corresponding to the model determined by the intrinsic mode function (IMFs) is solved according to the following formula (20), where T s is the sampling time, used to convert continuous time parameters into discrete time parameters, and λ1() is the first complex eigenvalue associated with the dynamic matrix A.

[0119]

[0120] The w calculated by formula (20) n Determined as the first-order natural frequency.

[0121] In the above scheme, the N4SID subspace identification technology is used to analyze the intrinsic mode function to determine the natural frequency corresponding to the model determined by the intrinsic mode function, including the first-order natural frequency of the wheel-rail.

[0122] In some embodiments, after obtaining the vertical vibration acceleration data of the train axle box on the target track along the time series in step S10, the method for identifying the high-frequency dynamic stiffness of the rail fasteners provided in the embodiment of the present application also includes: filtering and normalizing the vertical vibration acceleration data.

[0123] In the above scheme, in order to ensure that the vertical vibration acceleration data along the time series is smoother and more stable and to improve the data quality, the data is filtered and normalized.

[0124] In some embodiments, the design process of a method for identifying high-frequency dynamic stiffness of a rail fastener provided by the present application is as follows: Figure 3 As shown. This method does not require any additional equipment to be added to the existing track structure, has no destructive impact on the track, does not occupy window time, and eliminates the need for on-site monitoring by maintenance personnel, and does not have any additional impact on the operation of existing trains. Compared with existing methods, this method can identify the high-frequency dynamic stiffness of fasteners in real time when trains pass through various parts of the line in the operating line, and use high-frequency dynamic stiffness data to evaluate whether the vibration generated by the train during driving exceeds the standard, and guide the replacement time of the fastener elastic element when the vibration reduction fails, which is of great significance to the green and healthy operation of trains.

[0125] In addition, in existing patents, on-site identification of fastener stiffness is often limited to the measurement of a few fasteners on the entire line, which limits the breadth and efficiency of monitoring to a certain extent. The high-frequency dynamic stiffness identification method proposed in the embodiment of the present application realizes the rapid on-site identification of the high-frequency dynamic stiffness of fasteners at all positions in the entire track line, and can monitor the high-frequency dynamic stiffness of fasteners on the entire line in real time and uninterruptedly. It is not only efficient, but also not affected by environmental factors, providing a reliable solution for long-term and continuous monitoring of high-frequency dynamic stiffness of tracks.

[0126] For ease of understanding, the following data is taken as an example to identify the high-frequency dynamic stiffness of the fastener using the method provided in the embodiment of the present application:

[0127] like Figure 4 As shown, Figure 4This is a set of vertical vibration acceleration data of axle boxes of a certain city's subway tested at around 8 am. The train speed is 110km / h. The fastener system of this section line uses A-type fastener pads. Figure 4 The red dotted frame part is the axle box vertical vibration acceleration data measured when the train runs on the ordinary integral track section.

[0128] To maintain data stability, Figure 4 In the red dotted box, 20 data segments are randomly sampled, and the duration of each data segment is about 3 to 4 seconds. Taking one set of data as an example, the processing flow of the measured data is briefly introduced. Figure 5 As shown in Figure 6 shown.

[0129] Then, through empirical mode decomposition, the first-order intrinsic mode function (IMF1) of the wheel / rail is extracted from the processed axle box vertical vibration acceleration data, as shown in Figure 7 shown.

[0130] Then, the extracted intrinsic mode function (IMF1) is identified by N4SID subspace identification technology, and the first-order natural frequency corresponding to the intrinsic mode function (IMF1) is 57.9 Hz. Finally, the first-order natural frequency of the wheel-rail is input into the unsprung mass-track system analysis model. The specific parameters of the unsprung mass-track system analysis model are shown in Table 1. The high-frequency dynamic stiffness of the track fastener is identified to be 43.90 kN / mm.

[0131] parameter Numeric mr(kg / m) 60 L(m) 100 <![CDATA[EI(N·m 2 )]]> <![CDATA[6.62×10 5 ]]> <![CDATA[M U (kg)]]> 1473 / 2

[0132] Table 1 Specific parameters of the unsprung mass-track system analysis model

[0133] The same method and process as above are used to analyze the remaining 19 groups of measured axle box acceleration data. After the analysis is completed, the identification results of the first-order natural frequency of the wheel and rail are as follows: Figure 8 As shown in Figure 2, the first-order natural frequency of the wheel / rail is in the range of 54 to 60 Hz, and the corresponding high-frequency dynamic stiffness identification results are shown in Figure 2. Fig. 9 shown.

[0134] In order to verify the accuracy of the identification results, it is necessary to refer to the existing patent "Method and Device for Measuring Parameters of Rubber Pads for Tracks" (ZL201610057354.9) to conduct high-frequency dynamic stiffness tests on on-site fasteners. Since the on-site data was taken during the morning rush hour, the test preload was selected at 60kN. Under this preload condition, the high-frequency dynamic stiffness value of the on-site fasteners at 50Hz is 40.72kN / mm, and the high-frequency dynamic stiffness value at 60Hz is 41.22kN / mm. The test results at 50Hz and 60Hz only differ by 0.5kN / mm, and most of the identification frequencies are close to 60Hz. Therefore, the indoor stiffness test result of 60Hz, 41.22kN / mm, is selected as the benchmark value.

[0135] The high frequency dynamic stiffness of the fastener identified by the above field test is compared and analyzed with the reference value to verify the practicality of the method provided by the embodiment of the present application. Fig.10 As shown, 90% of the field test identification results fall within the error range of ±5kN / mm of the indoor test benchmark value. These data show that the method provided in the embodiment of the present application can effectively identify the high-frequency dynamic stiffness of the ordinary integral track bed track fasteners on site by analyzing the vertical vibration acceleration data of the train axle box, and has good practicality and reliability.

[0136] Fig.11 A schematic diagram of the structure of an electronic device provided in an embodiment of the present application, such as Fig.11 As shown, the electronic device includes: a processor 100, a memory 200 and a bus 300, wherein the processor 100 and the memory 200 communicate with each other through the bus 300; the memory 200 stores program instructions that can be executed by the processor 100, and the processor 100 calls the program instructions to execute the methods provided by the above-mentioned method embodiments.

[0137] Among them, the processor 100 includes one or more (only one is shown in the figure), which can be an integrated circuit chip with signal processing capabilities. The above-mentioned processor 100 can be a general-purpose processor 100, including a central processing unit 100 (Central Processing Unit, referred to as CPU), a micro control unit (Micro Controller Unit, referred to as MCU), a network processor 100 (Network Processor, referred to as NP) or other conventional processors 100; it can also be a special-purpose processor 100, including a neural network processor 100 (Neural-network Processing Unit, referred to as NPU), a graphics processor 100 (Graphics Processing Unit, referred to as GPU), a digital signal processor 100 (Digital Signal Processor, referred to as DSP), an application-specific integrated circuit (Application Specific Integrated Circuits, referred to as ASIC), a field programmable gate array (Field Programmable Gate Array, referred to as FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components. In addition, when there are multiple processors 100, some of them can be general-purpose processors 100, and the other part can be special-purpose processors 100.

[0138] The memory 200 includes one or more (only one is shown in the figure), which may be, but not limited to, a random access memory (RAM), a read only memory (ROM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), an electrically erasable programmable read-only memory (EEPROM), etc. The processor 100 and other possible components can access the memory 200 and read and / or write data therein.

[0139] In particular, one or more computer program instructions may be stored in the memory 200, and the processor 100 may read and execute these computer program instructions to implement the weak password scanning behavior identification method provided in the embodiment of the present application.

[0140] The bus 300 includes one or more (only one is shown in the figure) devices that can be used to communicate directly or indirectly with other devices to exchange data. Fig.11 The intermediate network device in the bus 300 can communicate with the server device and the client device. The bus 300 may include devices for wired and wireless communication, such as optical fiber, serial peripheral interface (SPI) module, integrated circuit bus (I2C), etc., and may also include devices for wireless communication, such as Bluetooth module, Wi-Fi module, mobile communication module (such as 4G, 5G module), etc.

[0141] Understandably, Fig.11 The structure shown is for illustration only. The electronic device may also include Fig.11 More or fewer components as shown, or with Fig.11 Different structures are shown. Fig.11 The components shown in the figure can be implemented by hardware, software or a combination thereof. The electronic device may be a physical device, such as a switch, a router, a server, a PC, etc., or a virtual device, such as a virtual machine, a virtualized container, etc. Moreover, the electronic device is not limited to a single device, but may also be a combination of multiple devices or a cluster consisting of a large number of devices.

[0142] This embodiment discloses a computer program product, including computer program instructions. When the computer program instructions are read and executed by a processor, the methods provided by the above method embodiments are executed.

[0143] This embodiment provides a computer-readable storage medium, wherein the computer-readable storage medium stores computer instructions, and the computer instructions enable the computer to execute the methods provided by the above-mentioned method embodiments.

[0144] The present application provides a device for identifying high-frequency dynamic stiffness of a rail fastener, comprising:

[0145] An acquisition module is used to acquire vertical vibration acceleration data of a train axle box on a target track along a time series, and to acquire an unsprung mass-track system analysis model, wherein the unsprung mass-track system analysis model is used to analyze the dynamic performance of the target track based on a Winkler foundation model, including a transcendental equation with a support stiffness per unit length of rail as an unknown parameter.

[0146] The first-order natural frequency determination module is used to determine the first-order natural frequency of the wheel-rail corresponding to the target track according to the vertical vibration acceleration data.

[0147] The high-frequency dynamic stiffness identification module is used to input the first-order natural frequency into the unsprung mass-track system analysis model, solve the transcendental equation to obtain the support stiffness, and identify the high-frequency dynamic stiffness of the track fastener corresponding to the target track according to the support stiffness.

[0148] The above description is only an embodiment of the present application and is not intended to limit the protection scope of the present application. For those skilled in the art, the present application may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included in the protection scope of the present application.

Claims

1. A method for identifying high frequency dynamic stiffness of rail fasteners, characterized in that: include: Obtaining vertical vibration acceleration data of axle boxes of trains on a target track along a time series, and obtaining an unsprung mass-track system analysis model, wherein the unsprung mass-track system analysis model is used to analyze the dynamic performance of the target track based on a Winkler foundation model, and includes a transcendental equation with a support stiffness per unit length of rail as an unknown parameter; Determining the first-order natural frequency of the wheel-rail corresponding to the target track according to the vertical vibration acceleration data; The first-order natural frequency is input into the unsprung mass-track system analysis model, the transcendental equation is solved to obtain the support stiffness, and the high-frequency dynamic stiffness of the track fastener corresponding to the target track is identified based on the support stiffness.

2. The method according to claim 1, characterized in that: Before obtaining the unsprung mass-track system analysis model, the method further includes: Constructing a vibration differential equation of the target track, wherein the vibration differential equation represents the dynamic performance of the rail and the wheel during vibration when the train wheel load acts on any point on the target track; According to the vibration differential equation, a vibration mode function of the rail is obtained; The transcendental equation is obtained according to the natural frequency characteristic equation of the vibration mode function.

3. The method according to claim 1, characterized in that The expression of the transcendental equation is: Among them, M U is half of the wheelset mass, k H is the linearized wheel-rail contact stiffness, EI is the bending stiffness of the rail, k is the support stiffness, m r is the mass of the rail per unit length, L is the length of the rail, α 4 =(λ 4 EI+k) / m r , where α 2 is the first-order natural frequency.

4. The method according to claim 1, characterized in that: Determining the first-order natural frequency of the wheel-rail corresponding to the target track according to the vertical vibration acceleration data includes: Using an empirical mode decomposition method, all inherent mode functions are extracted from the vertical vibration acceleration data; The N4SID subspace identification method is adopted to determine the state space model corresponding to the intrinsic mode function, and the first-order natural frequency is determined from the state space model.

5. The method according to claim 4, characterized in that The empirical mode decomposition method is used to extract all the intrinsic mode functions from the vertical vibration acceleration data, including: Identifying local maximum values ​​and local minimum values ​​in the vertical vibration acceleration data, and constructing an envelope between the local maximum values ​​and the local minimum values; extracting a first natural mode function from the vertical vibration acceleration data according to the envelope, and subtracting the first natural mode function from the vertical vibration acceleration data to obtain residual data; Repeat the following steps until the remaining data contains no more than two extreme values: A new intrinsic mode function is extracted from the residual data, and the new intrinsic mode function is subtracted from the residual data to obtain new residual data.

6. The method according to claim 5, characterized in that The adopting of the N4SID subspace identification method to determine the state space model corresponding to the intrinsic mode function, and determining the first-order natural frequency from the state space model, comprises: Constructing a past output matrix and a future output matrix based on the intrinsic mode function, wherein the past output matrix represents the output data of the system in the past period of time, and the future output matrix represents the output data of the assumed system in the future period of time; Based on the past output matrix, calculating the spatial projection of the future output matrix using LQ decomposition; Calculating the future state vector using a singular value decomposition method according to the spatial projection; Determine the state space model according to the future state vector, and determine the dynamic matrix of the state space model by using the least square method; The first-order natural frequency is determined according to the dynamic matrix.

7. The method according to any one of claims 1 to 6, characterized in that: After obtaining the vertical vibration acceleration data of the train axle box on the target track along the time series, the method further includes: The vertical vibration acceleration data is filtered and normalized.

8. A computer program product, characterized in that The method comprises computer program instructions, and when the computer program instructions are read and executed by a processor, the method according to any one of claims 1 to 7 is executed.

9. An electronic device, characterized in that: include: A processor, a memory and a bus, wherein the processor and the memory communicate with each other via the bus; The memory stores program instructions executable by the processor, and the processor can execute the method according to any one of claims 1 to 7 by calling the program instructions.

10. A computer-readable storage medium, characterized in that: include: The computer-readable storage medium stores computer instructions, and the computer instructions enable the computer to execute the method according to any one of claims 1 to 7.

Citation Information

Patent Citations

  • Urban railway system plate type vibration damping track structure analysis model selection method

    CN105677985A

  • Track rigidity time-frequency measurement method based on P2 force vibration

    CN113124998A

  • Rail fastener parameter optimization method based on vehicle-track coupling theory

    CN113656899A

  • Method and device for calculating service safety of steel rail

    CN115892133A

  • Construction method of vehicle-track dynamic model considering buckling state of fastener elastic strip

    CN118194501A

Cited By

  • Self-adaptive ultrasonic spindle control method and system with force, temperature and vibration sensing function

    CN121070112A