Abnormal wheel-rail relationship detection system and method

Abstract

The present invention discloses an abnormal wheel-rail relationship detection system. Its driving vibration signal acquisition module is used to acquire the driving vibration signal of any measurement area along the rail transit to be measured; the signal decomposition module is used to decompose a plurality of intrinsic mode functions; the feature extraction module is used to calculate the kurtosis values of each intrinsic mode function; the abnormal signal recognition module is used to compare the kurtosis values of each intrinsic mode function with the kurtosis threshold respectively; the fault location module is used to locate the time point of the abnormal fluctuation in the time domain of the driving vibration signal of the abnormal measurement area according to the time point of the extreme value of the abnormal intrinsic mode function in the time domain, so as to locate the position of the bogie with abnormal wheel-rail relationship in the train. The present invention realizes the accurate and rapid location of the bogie with abnormal wheel-rail relationship.

Description

Technical Field

[0001] The present invention relates to the technical field of optical cable detection, and in particular to a system and method for detecting abnormal wheel-rail relationship. Background Art

[0002] Delivering passengers and materials to their destinations safely and quickly is the primary task of rail transit. Among the many factors that affect the safe operation of trains, the wheel-rail relationship is an extremely important aspect. Abnormal wheel-rail relationship may cause train derailment, rollover and other accidents, which will bring serious safety hazards to rail transit. Periodic lathe processing of train wheels is the main method to ensure the normal wheel-rail relationship, but this method has a long cycle and cannot detect trains with abnormal wheel-rail relationship in time.

[0003] In order to ensure the safe and stable operation of trains and to promptly detect trains with abnormal wheel-rail relationships, researchers have proposed a series of methods for detecting abnormal wheel-rail relationships. For example, a camera is installed under the train to capture images of the track, thereby realizing the identification of abnormal fasteners. However, the cost of deploying cameras on each train is very high, and using images to identify faults can only detect defects on the surface of the wheel and rail. Some researchers have also deployed 10 pairs of FBG sensors on the track to detect wheel defects and analyzed the reasonable layout of the sensors. However, the coverage of 10 pairs of sensors is small, which is not conducive to the large-scale reuse of the sensing system. Some researchers have also installed a microphone near the bogie to collect the noise when the vehicle is running, and realized the classification of normal background noise and abnormal impact sound. However, the environment has a great influence on the effect of the microphone collecting sound, and it is also costly to deploy high-quality microphones on each vehicle. Summary of the invention

[0004] The purpose of the present invention is to provide a system and method for detecting an abnormal wheel-rail relationship, and the present invention realizes accurate and rapid positioning of a bogie with an abnormal wheel-rail relationship.

[0005] To achieve this purpose, the abnormal wheel-rail relationship detection system designed by the present invention includes a vehicle vibration signal acquisition module, a signal decomposition module, a feature extraction module, an abnormal signal recognition module and a fault location module; each vehicle vibration signal measurement area in the track distributed acoustic wave sensor system can sense the vehicle vibration signal of the corresponding measurement area;

[0006] The vehicle vibration signal acquisition module is used to acquire the vehicle vibration signal of any measurement area along the rail transit line to be measured;

[0007] The signal decomposition module is used to perform modal decomposition on the driving vibration signal in the measurement area using the variational modal decomposition method, thereby decomposing multiple intrinsic mode functions;

[0008] The feature extraction module is used to calculate the kurtosis values of each intrinsic mode function;

[0009] The abnormal signal recognition module is used to compare the kurtosis values of each intrinsic mode function with the kurtosis threshold respectively. If the kurtosis value of one intrinsic mode function is greater than the kurtosis threshold, then the intrinsic mode function with the kurtosis value greater than the kurtosis threshold is the abnormal intrinsic mode function, and the original train running vibration signal corresponding to the intrinsic mode function with the kurtosis value greater than the kurtosis threshold is the abnormal train running vibration signal of the measurement area;

[0010] The fault location module is used to locate the time point of the abnormal fluctuation in the time domain of the abnormal train running vibration signal in the measurement area according to the time point of the extreme value of the abnormal intrinsic mode function in the time domain, so as to locate the position of the bogie with abnormal wheel-rail relationship in the train.

[0011] The beneficial effects of the present invention:

[0012] First, the present invention obtains the vibration signal (train running vibration signal) caused by the train running vibration through the DAS system, and then uses variational mode decomposition (VMD) to decompose the vibration signal. Then, the kurtosis values of each mode obtained by the signal decomposition are calculated respectively, and the kurtosis threshold is combined to distinguish the normal train running vibration signal from the abnormal train running vibration signal. Finally, the present invention realizes the location of the bogie with abnormal wheel-rail relationship based on the extreme value position in the abnormal mode.

[0013] The present invention utilizes the fiber optic distributed acoustic sensing (DAS) system, which can continuously monitor the vibration signals within a range of dozens of kilometers, and has the advantages of high sensitivity, high sampling rate, strong anti-electromagnetic interference ability, low cost, and easy implementation of large-scale multiplexing, etc., to realize full-time and full-domain safety monitoring. It improves the reliability of the operation of rail transit trains and avoids accidents such as derailment and rollover of trains caused by abnormal wheel-rail relationships. Description of the Drawings

[0014] Figure 1 is the schematic diagram of the principle of the present invention;

[0015] Figure 2 is the basic schematic diagram of the DAS system used in the present invention;

[0016] Figure 3 is the waveform diagram of the typical train running vibration signal sensed by the DAS system;

[0017] Figure 4 is the comparison diagram of the normal train running vibration signal and the abnormal train running vibration signal;

[0018] Figure 5 is the VMD decomposition result diagram of the normal train running vibration signal;

[0019] Figure 6 This is the VMD decomposition result diagram of abnormal train running vibration signal;

[0020] Figure 7 It is a comparison chart of the kurtosis eigenvectors of normal signals and abnormal signals;

[0021] Figure 8 It is a comparison diagram of abnormal mode and abnormal signal of abnormal signal;

[0022] Figure 9 The abnormal driving voting map for the first day among four days;

[0023] Figure 10 The abnormal driving voting map for the second day among four days;

[0024] Figure 11 The voting map for abnormal driving on the third day among four days;

[0025] Figure 12 This is the abnormal driving voting map for the 4th day among 4 days. DETAILED DESCRIPTION

[0026] The present invention is further described in detail below with reference to the accompanying drawings and specific embodiments:

[0027] like Figure 1 The abnormal wheel-rail relationship detection system shown in the figure includes a vehicle vibration signal acquisition module, a signal decomposition module, a feature extraction module, an abnormal signal recognition module and a fault location module; two adjacent ultra-weak fiber Bragg gratings and the optical fiber therebetween in the rail transit distributed acoustic wave sensing system constitute a vehicle vibration signal measurement area, and each vehicle vibration signal measurement area can sense the vehicle vibration signal of the corresponding measurement area;

[0028] The vehicle vibration signal acquisition module is used to acquire the vehicle vibration signal of any measurement area along the rail transit line to be measured;

[0029] The signal decomposition module is used to perform modal decomposition on the driving vibration signal in the measurement area using the variational modal decomposition method, thereby decomposing multiple intrinsic mode functions;

[0030] The feature extraction module is used to calculate the kurtosis value of each intrinsic mode function. The kurtosis can measure the prominence of abnormal fluctuations in the signal. The kurtosis is a statistical measure used to describe the time domain distribution of a time series. It can indicate the peak and impulse of the time series. Compared with the Gaussian distribution, it can measure the shape characteristics of the time series. Specifically, when there are extreme values in the time series, the corresponding kurtosis will increase significantly.

[0031] The abnormal signal recognition module is used to compare the kurtosis values of each intrinsic mode function with the kurtosis threshold respectively. If the kurtosis value of one intrinsic mode function is greater than the kurtosis threshold, then the intrinsic mode function with the kurtosis value greater than the kurtosis threshold is the abnormal intrinsic mode function, and the original train-running vibration signal corresponding to the intrinsic mode function with the kurtosis value greater than the kurtosis threshold is the abnormal train-running vibration signal of the measurement area;

[0032] The fault location module is used to locate the time point in the time domain where the abnormal fluctuation occurs in the abnormal train-running vibration signal of the measurement area according to the time point in the time domain where the extreme value of the abnormal intrinsic mode function is located, so as to locate the position of the bogie with abnormal wheel-rail relationship in the train.

[0033] The sensing system used in a certain transportation line is a distributed acoustic sensing (DAS) system. As Figure 2 shown, in the system, the continuous light emitted by the narrow linewidth laser is modulated into an optical pulse sequence after passing through the electro-optic modulator. Then, the optical pulse sequence is amplified by the erbium-doped fiber amplifier and finally incident on the ultra-weak fiber Bragg grating (UWFBG) array. The amplified pulsed light enters the unbalanced Michelson interferometer after being reflected by the UWFBG array. The interferometer restores the amplitude of the time-domain train-running vibration signal by demodulating the phase change caused by the optical length change in the fiber between two adjacent UWFBGs, and outputs the interfered result to three photodetectors. In this way, the time-domain train-running vibration signal can be obtained after photoelectric conversion. Since the length of the delay fiber in the Michelson interferometer is the same as the distance between adjacent UWFBGs in the optical cable, every two adjacent UWFBGs and the fiber between them form a sensor, that is, a measurement area. In rail transit, the vibration signals occurring in this interval can be sensed, and according to needs, the vibration signals when the train passes by can be intercepted, that is, the train-running vibration signals;

[0034] The train-running vibration signal acquisition module combines Figure 2 the above-mentioned sensing system to acquire the train-running vibration signals of each measurement area along the rail transit to be measured. The typical train-running vibration signals sensed by the module are as Figure 3 shown. The sensed signals can reflect the structural characteristics of the rail transit vehicle, such as the number of carriages and bogies. In Figure 3 it, the positions of six carriages are marked with C1, C2, C3, C4, C5, and C6 respectively. And each carriage contains 2 bogies;

[0035] In the above technical solution, the comparison chart of the typical normal train-running vibration signal and the abnormal train-running vibration signal sensed by the DAS system is as Figure 4As shown, the abnormal train running vibration signal has abnormal fluctuations in the time domain, and the amplitude of the abnormal fluctuations is not prominent relative to the main part of the signal. Therefore, the abnormal signal cannot be directly distinguished by amplitude. In addition, the frequency range where the abnormal fluctuations of the abnormal signal are located is greater than 20 Hz, while the frequency of the main component of the signal is less than 15 Hz. In contrast, the time-domain waveform of the normal signal is smooth and shows regular fluctuations, and the main frequency component of the entire signal is below 15 Hz. Therefore, the variational mode decomposition method can separate the abnormal fluctuations in the signal into the modes. And, the 4-mode VMD decomposition result of the normal signal is as Figure 5 shown, and the 4-mode VMD decomposition result of the abnormal signal is as Figure 6 shown. In Figure 6 it can also be found that the abnormal fluctuations in the abnormal signal are mainly decomposed into mode 3.

[0036] In the above technical solution, the abnormal signal recognition module based on kurtosis threshold is used to distinguish the two signals using the kurtosis threshold. This is because the element values in the kurtosis feature vector of the normal signal are small, and the kurtosis value of the mode containing abnormal fluctuations obtained by decomposing the abnormal signal is large. As shown in the feature vector comparison diagram in Figure 7 , the kurtosis value of mode 3 of the abnormal signal is greater than 10, while the kurtosis values of each mode of the normal signal are less than 10. Therefore, the kurtosis value of 10 is selected as the threshold for distinguishing the normal signal and the abnormal signal. When a mode with a kurtosis greater than 10 is detected, the corresponding mode is considered an abnormal mode, and the corresponding original signal is considered an abnormal signal. The kurtosis of the VMD decomposition result is as Figure 7 shown. Obviously, the kurtosis value of mode 3 of the abnormal signal is much greater than the kurtosis values of each mode of the normal signal.

[0037] In the above technical solution, the fault location module based on the extreme position of the abnormal mode is used to locate the position of the abnormal fluctuation in the original signal using the position of the extreme point of the mode. And the wheel-rail vibration signal is closely related to the structure of the train. The position of the abnormal fluctuation in the wheel-rail vibration signal can indicate the position of the bogie with abnormal wheel-rail relationship. After passing through the abnormal signal recognition module based on kurtosis threshold, the abnormal signal and its abnormal mode recognized by this module can be obtained. And the position of the extreme value in the abnormal mode is consistent with the position where the abnormal fluctuation in the abnormal signal is located. Therefore, the position of the abnormal fluctuation in the abnormal signal can be located through the extreme position of the abnormal mode, so as to locate the position of the bogie with abnormal wheel-rail relationship in the train. The comparison diagram of the abnormal signal and the abnormal mode of the above typical abnormal signal is as Figure 8As shown, in the comparison chart, the position of the extreme point of the abnormal mode, i.e., mode 3, coincides with the abnormal fluctuation position of the abnormal signal, indicating an abnormal wheel-rail relationship in the first bogie of the train. Moreover, during subsequent train maintenance, it was found that the wheels of the first bogie of the train corresponding to this driving vibration signal had problems with tread wear.

[0038] In the above technical solution, it further includes an abnormal driving voting chart drawing module. The abnormal driving voting chart drawing module is used to calculate the number of times that the driving vibration signals sensed by each corresponding measurement area of the driving vibration signals sensed by each measurement area of the track distributed acoustic sensing system after a single train travels through all the measurement areas are identified as abnormal measurement area driving vibration signals using the kurtosis threshold method. The abnormal driving voting chart drawing module draws an abnormal driving voting chart based on the number of abnormal measurement area driving vibration signals. The abscissa of the abnormal driving voting chart is the train driving serial number, and the ordinate is the number of measurement areas that sense abnormal measurement area driving vibration signals. In a certain traffic line, 70 measurement areas were selected for voting, and the driving vibration signals of the train throughout four days were used respectively, and thus the abnormal driving voting charts as shown in Figure 9 , Figure 10 , Figure 11 , Figure 12 were obtained. In the voting chart of each day, the train drivings with a higher number of voting measurement areas form multiple arithmetic progressions, which is consistent with the train scheduling arrangement in this traffic line. And the train drivings marked by voting are consistent with the abnormal trains found in actual maintenance.

[0039] In the above technical solution, the signal decomposition module uses the variational mode decomposition method to perform mode decomposition on the driving vibration signal of this measurement area. The specific method for decomposing multiple intrinsic mode functions is as follows:

[0040] Solve the constrained variational problem:

[0041]

[0042] where, {u k} is defined as the set of the first mode to the Kth mode {u k}:={u1,...,u K}, {ω k} is defined as the set of the center frequency of the first mode to the center frequency of the Kth mode {ω k}:={ω1,...,ω K}, u k (t) is defined as the kth mode, ω k is defined as the center frequency of the kth mode, the value range of k is 1≤k≤K, K is the decomposition number set by the variational mode decomposition, δ(t) represents the impulse signal, j is the imaginary part unit in the complex number, t is the time of the signal, and e is the natural constant. Denote the square of the gradient as L 2 norm, denote the value of B when the minimum value of formula A is taken, and x(t) is the driving vibration signal of the measuring area;

[0043] Among them, the constrained variational problem is constructed in the following way: First, obtain the unilateral spectrum of the mode through Hilbert transform. Second, transform the spectrum of each mode to the baseband. Third, obtain the bandwidth of each mode by using the demodulated signal. Combining the above three steps, the constrained variational problem actually hopes that the sum of the bandwidths of the decomposed modes is the smallest;

[0044] Based on the Alternating Direction Method of Multipliers (ADMM), the mode u k , the center frequency ω k and the update formulas of the Lagrange multiplier λ can be obtained:

[0045]

[0046]

[0047]

[0048] Among them, α is the quadratic penalty parameter, λ is a function of the Lagrange multiplier, n is the number of iterations, ^ represents the Fourier transform, τ is the penalty parameter, ω is the frequency, and x is the driving vibration signal of the measuring area in the update formula, represents the result of the (n + 1)-th iteration of the k-th mode in the frequency domain, represents the Fourier transform result of the driving vibration signal of the measuring area, represents the iterative result of the i-th mode in the frequency domain. When i < k, this iterative result is the result of the (n + 1)-th iteration. When i > k, this iterative result is the result of the n-th iteration, represents the Fourier transform result of the Lagrange multiplier function, represents the result of the center frequency of the k-th mode at the (n + 1)-th iteration, represents the Fourier transform result of the k-th mode, λ n+1 represents the result of the (n + 1)-th iteration of the Lagrange multiplier function, λ n represents the result of the n-th iteration of the Lagrange multiplier function, u k n+1 represents the result of the (n + 1)-th iteration of the k-th mode;

[0049] After obtaining the update formulas of the mode u k , the center frequency ω k and the function λ of the Lagrange multiplier, iteration can be started by initializing the mode, the center frequency, and the Lagrange multiplier function. In each iteration, first use

[0050]

[0051] Update the first mode to the K-th mode in sequence, and then use

[0052]

[0053] Update the center frequency of the first mode to the center frequency of the K-th mode in sequence, and then use

[0054]

[0055] Update the Lagrange multiplier function. Through iteration, finally when the set convergence tolerance or the upper limit of iteration is reached, the first mode to the K-th mode can be obtained. Finally, transform each mode from the frequency domain to the time domain, and the process of variational mode decomposition to decompose the signal into modes is completed;

[0056] In this process, each intrinsic mode function obtained by decomposition is defined as an amplitude-modulated frequency-modulated (AM-FM) signal, and the form is as follows:

[0057] u k (t) = A k (t)cos(φ k (t))

[0058] where, φ k (t) is a non-decreasing phase function in the intrinsic mode function, φ k ′(t) ≥ 0, A k (t) is a non-negative envelope function in the intrinsic mode function, A k (t) ≥ 0, and the instantaneous frequency function in the intrinsic mode function is defined as the derivative ω k (t): = φ k ′(t), and the change rate of the instantaneous frequency function and the change rate of the envelope function are both slower than the change rate of the phase function;

[0059] In addition, there is the following relationship between the intrinsic mode function obtained by variational mode decomposition and the driving vibration signal in the measurement area:

[0060]

[0061] where, x(t) is the driving vibration signal in the measurement area, u k (t) represents the k-th intrinsic mode function, and K is the decomposition number of variational mode decomposition.

[0062] In the above technical solution, the specific method for the feature extraction module to calculate the kurtosis value of each intrinsic mode function is:

[0063]

[0064] Among them, V k is the kurtosis value of the k-th intrinsic mode function, μ and σ represent the mean and standard deviation of the intrinsic mode function, E{·} represents the expectation, and IMF k represents the k-th intrinsic mode function.

[0065] In the above technical solution, the specific method for the fault location module to locate the time point of the abnormal fluctuation in the time domain of the train running vibration signal in the abnormal measurement area according to the time point of the extreme value of the abnormal intrinsic mode function in the time domain is as follows: The time point of the extreme value of the abnormal intrinsic mode function in the time domain coincides with the time point of the abnormal fluctuation in the time domain of the train running vibration signal in the abnormal measurement area. By locating the time point of the extreme value of the abnormal intrinsic mode function, the time point of the abnormal fluctuation in the train running vibration signal in the abnormal measurement area can be located.

[0066] In the above technical solution, the specific method for the fault location module to locate the bogie position with abnormal wheel-rail relationship in the train according to the time point of the abnormal fluctuation in the time domain of the train running vibration signal in the abnormal measurement area is as follows: The train running vibration signal in the measurement area is generated when the train passes through the measurement area. From the front to the end of the train, the first carriage passes through the measurement area first, and the last carriage passes through the measurement area last. For each carriage, the first bogie in the carriage passes through the measurement area first, and the second bogie passes through the measurement area later. The difference in the time points in the train running vibration signal in the measurement area corresponds to different bogies of the train. Therefore, the position with abnormal wheel-rail relationship in the train structure can be located through the time point of the abnormal fluctuation in the train running vibration signal in the abnormal measurement area, that is, the bogie position with abnormal wheel-rail relationship can be located.

[0067] In the above technical solution, the number of intrinsic mode functions decomposed by the signal decomposition module is determined by the number of modes decomposed by the variational mode decomposition method.

[0068] In the above technical solution, the abnormal signal recognition module compares the kurtosis values of each intrinsic mode function with the kurtosis threshold respectively. If the kurtosis values of all intrinsic mode functions are less than or equal to the kurtosis threshold, then all intrinsic mode functions are normal intrinsic mode functions, and the original train running vibration signal corresponding to the intrinsic mode function is a normal train running vibration signal in the measurement area.

[0069] An abnormal wheel-rail relationship detection method includes the following steps:

[0070] Step 1: Obtain the train running vibration signal of any measurement area along the rail transit to be measured;

[0071] Step 2: Use the variational mode decomposition method to perform modal decomposition on the train running vibration signal in this measurement area, so as to decompose multiple intrinsic mode functions;

[0072] Step 3: Calculate the kurtosis values of each intrinsic mode function;

[0073] Step 4: Compare the kurtosis values of each intrinsic mode function with the kurtosis threshold respectively. If the kurtosis value of one intrinsic mode function is greater than the kurtosis threshold, then the intrinsic mode function with the kurtosis value greater than the kurtosis threshold is the abnormal intrinsic mode function, and the original train running vibration signal corresponding to the intrinsic mode function with the kurtosis value greater than the kurtosis threshold is the abnormal train running vibration signal of the measurement area;

[0074] Step 5: Locate the time points in the time domain where the abnormal fluctuations in the abnormal train running vibration signal of the measurement area are located according to the time points in the time domain where the extreme values of the abnormal intrinsic mode function are located, so as to locate the position of the bogie with abnormal wheel-rail relationship in the train.

[0075] The content not described in detail in this specification belongs to the prior art well known to those skilled in the art.

Claims

1. An abnormal wheel-rail relationship detection system, characterized in that: It includes a train running vibration signal acquisition module, a signal decomposition module, a feature extraction module, an abnormal signal recognition module, and a fault location module; each train running vibration signal measurement area in the track distributed acoustic sensing system can sense the train running vibration signal of the corresponding measurement area; The train running vibration signal acquisition module is used to acquire the train running vibration signal of any one measurement area along the rail transit to be measured; The signal decomposition module is used to perform modal decomposition on the train running vibration signal of this measurement area by using the variational mode decomposition method, so as to decompose multiple intrinsic mode functions; The feature extraction module is used to calculate the kurtosis values of each intrinsic mode function; The abnormal signal recognition module is used to compare the kurtosis values of each intrinsic mode function with the kurtosis threshold respectively. If the kurtosis value of one intrinsic mode function is greater than the kurtosis threshold, then the intrinsic mode function greater than the kurtosis threshold is the abnormal intrinsic mode function, and the original train running vibration signal of the measurement area corresponding to the intrinsic mode function greater than the kurtosis threshold is the abnormal train running vibration signal of the measurement area; The fault location module is used to locate the time point in the time domain where the abnormal fluctuation in the abnormal train running vibration signal of the measurement area is located according to the time point in the time domain where the extreme value of the abnormal intrinsic mode function is located, so as to locate the position of the bogie with abnormal wheel-rail relationship in the train.

2. The abnormal wheel-rail relationship detection system according to claim 1, wherein: It further includes an abnormal train voting map drawing module, and the abnormal train voting map drawing module is used to calculate the number of times that the train running vibration signal of each corresponding measurement area sensed by each train running vibration signal measurement area is identified as the abnormal train running vibration signal of the measurement area by using the kurtosis threshold method after a single train passes through all the train running vibration signal measurement areas in the track distributed acoustic sensing system.

3. The abnormal wheel-rail relationship detection system according to claim 2, characterized in that: The abnormal train voting map drawing module draws an abnormal train voting map according to the number of abnormal train running vibration signals of the measurement area. The abscissa of the abnormal train voting map is the train running serial number, and the ordinate is the number of measurement areas that sense the abnormal train running vibration signal of the measurement area.

4. The abnormal wheel-rail relationship detection system according to claim 1, wherein: The specific method for the signal decomposition module to perform modal decomposition on the train running vibration signal of this measurement area by using the variational mode decomposition method to decompose multiple intrinsic mode functions is as follows: Each intrinsic mode function is defined as an amplitude-modulated and frequency-modulated signal, and the form is as follows: u k u(t) = A k u(t) cos(φ k (t)) where, φ k (t) is a non-decreasing phase function in the intrinsic mode function, φ k ′(t) ≥ 0, A k (t) is a non-negative envelope function in the intrinsic mode function, A k (t) ≥ 0, and the instantaneous frequency function in the intrinsic mode function is defined as the derivative ω of the phase function k (t):=φ k ′(t), and the change rates of both the instantaneous frequency function and the envelope function are slower than the change rate of the phase function; In addition, there is the following relationship between the intrinsic mode function obtained by variational mode decomposition and the train running vibration signal of the measurement area: Among them, x(t) is the driving vibration signal of the measurement area, and u k (t) represents the k-th intrinsic mode function, and K is the decomposition number of variational mode decomposition.

5. The abnormal wheel-rail relationship detection system according to claim 1, characterized in that: The specific method for the feature extraction module to calculate the kurtosis values of each intrinsic mode function is as follows: where, V k is the kurtosis value of the k-th intrinsic mode function, μ and σ represent the mean and standard deviation of the intrinsic mode function, E{·} represents taking the expected value, and IMF k represents the k-th intrinsic mode function.

6. The abnormal wheel-rail relationship detection system according to claim 1, wherein: The specific method for the fault location module to locate the time point in the time domain where the abnormal fluctuation in the abnormal train running vibration signal of the measurement area is located according to the time point in the time domain where the extreme value of the abnormal intrinsic mode function is located is: the time point in the time domain where the extreme value of the abnormal intrinsic mode function is located coincides with the time point in the time domain where the abnormal fluctuation in the abnormal train running vibration signal of the measurement area is located. By locating the time point of the extreme value of the abnormal intrinsic mode function, the time point of the abnormal fluctuation in the abnormal train running vibration signal of the measurement area is located.

7. The abnormal wheel-rail relationship detection system according to claim 1, wherein: The specific method for the fault location module to locate the bogie position with abnormal wheel-rail relationship in the train according to the time point of the abnormal fluctuation in the time domain of the driving vibration signal in the abnormal measurement area is as follows: The driving vibration signal in the measurement area is generated when the train passes through the measurement area. From the head to the tail of the train, the first car passes through the measurement area first, and the last car passes through the measurement area last. For each car, the first bogie in the car passes through the measurement area first, and the second bogie passes through the measurement area later. The different time points in the driving vibration signal in the measurement area correspond to different bogies of the train. Therefore, the position with abnormal wheel-rail relationship in the train structure can be located through the time point of the abnormal fluctuation in the driving vibration signal in the abnormal measurement area, that is, the bogie position with abnormal wheel-rail relationship can be located.

8. The abnormal wheel-rail relationship detection system according to claim 1, characterized in that: The number of intrinsic mode functions decomposed by the signal decomposition module is determined by the number of modes decomposed by the variational mode decomposition method.

9. The abnormal wheel-rail relationship detection system according to claim 1, wherein: The abnormal signal recognition module compares the kurtosis values of each intrinsic mode function with the kurtosis threshold respectively. If the kurtosis values of all intrinsic mode functions are less than or equal to the kurtosis threshold, then all intrinsic mode functions are normal intrinsic mode functions, and the original driving vibration signal in the measurement area corresponding to the intrinsic mode function is a normal driving vibration signal in the measurement area.

10. A method for detecting abnormal wheel-rail relationship, characterized in that, It includes the following steps: Step 1: Obtain any driving vibration signal in the measurement area along the rail transit to be measured; Step 2: Use the variational mode decomposition method to perform mode decomposition on the driving vibration signal in the measurement area, so as to decompose multiple intrinsic mode functions; Step 3: Calculate the kurtosis values of each intrinsic mode function; Step 4: Compare the kurtosis values of each intrinsic mode function with the kurtosis threshold respectively. If the kurtosis value of one intrinsic mode function is greater than the kurtosis threshold, then the intrinsic mode function greater than the kurtosis threshold is an abnormal intrinsic mode function, and the original driving vibration signal in the measurement area corresponding to the intrinsic mode function greater than the kurtosis threshold is an abnormal driving vibration signal in the measurement area; Step 5: According to the time point of the extreme value of the abnormal intrinsic mode function in the time domain, locate the time point of the abnormal fluctuation in the abnormal driving vibration signal in the time domain, so as to locate the position of the bogie with abnormal wheel-rail relationship in the train.

Images