Railway point switch fault analysis method and system

Abstract

The invention discloses a railway switch machine fault analysis method and system. The method comprises the steps that three-phase action current is converted into three-phase action current digital signals according to a preset digital conversion strategy; cyclic frequency information in the three-phase action current digital signal is extracted, and an enhanced envelope spectrum is generated; quantifying and accumulating relative characteristics of a plurality of cyclic frequencies in the enhanced envelope spectrum to obtain a diagnosis characteristic spectrum, and identifying an equal-frequency interval harmonic structure in the enhanced envelope spectrum according to the diagnosis characteristic spectrum to obtain cyclic frequencies; and feeding the cyclic frequency to a preset maximum cyclostationary blind deconvolution, outputting the maximum cyclostationary blind deconvolution to obtain a repeated transient pulse caused by the local defect, and determining a fault signal of the railway switch machine according to the repeated transient pulse. The accuracy of railway point switch fault recognition is improved, fault related signals can be well recognized under the condition that strong noise interference exists outside, and troubleshooting of railway point switch faults is facilitated.

Description

Technical Field

[0001] The present invention belongs to the technical field of rail transit monitoring, and in particular relates to a railway switch machine fault analysis method and system. Background Art

[0002] Railway switch is an important equipment in the railway transportation system. It is mainly used to change the direction of railway turnouts to ensure that trains run along the predetermined track. The reliability of switch is crucial to the safety and efficiency of railway transportation. Therefore, it is of great significance to identify switch faults in time and repair them. With the continuous expansion of railway transportation, the problem of switch faults has gradually attracted attention. Switch faults may be caused by multiple factors such as electrical, mechanical, and environmental factors. Common faults include the inability of switch to open and close normally, inaccurate positioning, failure of the electronic control system, and sensor failure. These faults may not only cause danger during train operation, but also cause delays in railway operations and increase maintenance costs. At present, traditional switch fault detection methods mainly rely on manual inspections and regular inspections. However, manual inspections are labor-intensive, inefficient, and difficult to detect potential faults in time. Especially in the huge railway network, the coverage and accuracy of manual inspections are limited. The use of automated and intelligent fault detection technology has become an important means to improve the safety and operational efficiency of railway transportation.

[0003] Existing fault analysis of switch machines relies heavily on the operator's experience and intuition. The fault phenomenon often requires the operator to make a preliminary judgment through visual observation, testing, etc., and then conduct manual inspection. This method is not only prone to omissions or misjudgments, but also has limited efficiency and accuracy in manual judgment when facing complex faults. Summary of the invention

[0004] The present invention provides a railway switch machine fault analysis method and system, which are used to solve the technical problem of limited efficiency and accuracy of manual judgment when facing complex faults.

[0005] In a first aspect, the present invention provides a railway switch machine fault analysis method, comprising: Acquire the three-phase action current of the railway switch machine, and convert the three-phase action current into a three-phase action current digital signal according to a preset digital conversion strategy; Extracting the cycle frequency information in the three-phase action current digital signal and generating an enhanced envelope spectrum; quantifying and accumulating fault features of multiple cyclic frequencies in the enhanced envelope spectrum to obtain a diagnostic characteristic spectrum, and identifying an equi-frequency-interval harmonic structure in the enhanced envelope spectrum according to the diagnostic characteristic spectrum to obtain a cyclic frequency; The cyclic frequency is fed into a preset maximum cyclostationary blind deconvolution, the maximum cyclostationary blind deconvolution outputs a repetitive transient pulse caused by a local defect, and a fault signal of a railway switch is determined according to the repetitive transient pulse.

[0006] In a second aspect, the present invention provides a railway switch machine fault analysis system, comprising: An acquisition module is configured to acquire the three-phase action current of the railway switch machine and convert the three-phase action current into a three-phase action current digital signal according to a preset digital conversion strategy; A generating module configured to extract cyclostationarity information from the three-phase action current digital signal and generate an enhanced envelope spectrum; an identification module configured to quantify and accumulate fault features of multiple cyclic frequencies in the enhanced envelope spectrum to obtain a diagnostic feature spectrum, and identify an equi-frequency-interval harmonic structure in the enhanced envelope spectrum according to the diagnostic feature spectrum to obtain a cyclic frequency; The analysis module is configured to feed the cyclic frequency into a preset maximum cyclostationary blind deconvolution, the maximum cyclostationary blind deconvolution outputs a repetitive transient pulse caused by a local defect, and determines a fault signal of a railway switch according to the repetitive transient pulse.

[0007] According to a third aspect, an electronic device is provided, comprising: at least one processor, and a memory communicatively connected to the at least one processor, wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the steps of the railway switch fault analysis method of any embodiment of the present invention.

[0008] In a fourth aspect, the present invention further provides a computer-readable storage medium having a computer program stored thereon, wherein when the program instructions are executed by a processor, the processor executes the steps of the railway switch machine fault analysis method of any embodiment of the present invention.

[0009] The railway switch fault analysis method and system of the present application use a signal centralized monitoring system to collect the three-phase action current of the switch through a current sensor, extract the cyclic frequency information of the current digital signal and generate an enhanced envelope spectrum (EES), and obtain a diagnostic feature spectrum (DFS) by quantifying and accumulating the relative significance of multiple frequencies. The equi-frequency-interval harmonic structure (EIHS) in the EES can be identified, and an accurate cyclic frequency can be provided. The estimated cyclic frequency is fed into a maximum cyclostationary blind deconvolution (CYCBD) to retrieve repetitive transient pulses caused by local defects, thereby realizing adaptive diagnosis of railway switch faults, improving the accuracy of railway switch fault identification, and enabling fault-related signals to be well identified even in the presence of strong external noise interference, which is beneficial to the troubleshooting of railway switch faults. BRIEF DESCRIPTION OF THE DRAWINGS

[0010] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0011] Figure 1 A flow chart of a railway switch machine fault analysis method provided by one embodiment of the present invention; Figure 2 A structural block diagram of a railway switch machine fault analysis system provided by one embodiment of the present invention. DETAILED DESCRIPTION

[0012] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0013] See also Figure 1 , which shows a flow chart of a railway switch machine fault analysis method of the present application.

[0014] like Figure 1 As shown, the railway switch machine fault analysis method specifically includes the following steps: Step S101, obtaining a three-phase operating current of a railway switch machine, and converting the three-phase operating current into a three-phase operating current digital signal according to a preset digital conversion strategy.

[0015] In this step, the expression of the digital transformation strategy is: , , , , In the formula, is the digital conversion function of the characteristic factor of the lightweight current signal, is the partial discharge signal amplitude coefficient, is a partial discharge signal with unit amplitude, is a pure partial discharge signal, For delay, To detect signal noise, is the acquisition length, is the reference signal, is the effective value of the electrical signal, is the peak value of the electrical signal, is the electrical signal pulse width, is the signal duty cycle, is the main frequency component of the signal, is the signal harmonic distribution, is the signal spectrum energy entropy, For short-time overcurrent signal, For glitch signal, is the transient waveform distortion, is the signal activity, To reduce the sampling rate when the signal is stable, Trigger high sampling rate when a mutation occurs. is the similarity of adjacent sampling points, is the effective value error, For high temperature drift, For low temperature drift, is the environmental noise, It is a weak current signal. is the adaptive threshold, is the differential noise, For noise suppression, is the noise suppression weight, is the signal time information, is the partial discharge signal weight, is the electrical signal weight.

[0016] In this embodiment, a digital conversion strategy is adopted to compress and simplify the characteristic factors of the current signal so that it can be efficiently processed with fewer computing and storage resources after being digitized, helping the system to reduce the computing burden, power consumption, and optimize performance while maintaining accuracy. The appropriate conversion strategy can be flexibly selected according to the system requirements, resource limitations, and processing objectives.

[0017] Step S102, extracting the cycle frequency information in the three-phase action current digital signal and generating an enhanced envelope spectrum.

[0018] In this step, the three-phase action current digital signal is short-time Fourier transformed. The STFT can be expressed as: , in, , is the discrete frequency, is the sampling frequency, is the frequency resolution, is the window length, For time shift, is a window function, is the discrete frequency period, is the signal time-frequency, For time, is the frequency bandwidth, is the window interval; Scanning spectrum It can be defined as: , in, The corresponding frequency range is , is the frequency weight, For framing, Scan the spectrum for the previous frequency, For timeline information, is the sampling weight; The transformed three-phase action current digital signal is denoised according to a preset denoising processing strategy to obtain a target three-phase action current digital signal; The target three-phase action current digital signal is subjected to spectrum correlation analysis to obtain an enhanced envelope spectrum. The spectrum correlation analysis is specifically as follows: the signal is decomposed into different frequency components from the time domain, and the spectrum consistency of the signal in time can be quantified by correlating the spectrum of different time windows, and the cyclic frequency is extracted using fast spectrum correlation to generate the corresponding enhanced envelope spectrum.

[0019] It should be noted that the expression of the noise reduction processing strategy is: , , , , In the formula, For a stable target signal, is a non-stationary target signal, is additive noise, is a multi-component FM signal, is a non-convex regular term, is the L0 norm, is the signal sparsity, For time shift, is the frequency shift, To cover the signal, To tune the frequency space, is the quantity of the component, For the frequency modulation rule, is the dictionary redundancy, is the variable component, is super-Gaussian sparsity, To over-penalize large coefficients, is iterative reweighting, is the Lp norm, is the hard threshold, is the half threshold, is the pre-denoising value, To separate high frequency noise, is a low frequency signal, is the wavelet threshold, is the residual noise, To suppress the amount of impulse noise, is the noise standard deviation, is the number of dictionary atoms, are atomic parameters, is the proximal operator, is the initialization coefficient, is the time-frequency distribution, is additive Gaussian noise, As the impact component, is a transient non-stationary signal, For the wave group, For the ST segment, is the signal-to-noise ratio of the characteristic component, is a coarse-grained dictionary, is the compression coefficient vector, To incrementally update the weights, is the online calculation coefficient, For non-convex optimization of local minima, For high computational complexity, is the noise sensitivity, Automatically search for hyperparameters for Bayesian optimization, is a non-stationary target echo, is a high frequency transient signal, For high-precision time-frequency representation, is the optimization coefficient, It is a low-precision time-frequency representation.

[0020] In an embodiment, a noise reduction processing strategy extracts useful information from a complex signal, especially when the signal contains noise and has non-stationary characteristics (i.e., the statistical characteristics of the signal change over time). By selecting a suitable time-frequency analysis method according to the signal characteristics and combining it with noise suppression technology, the accuracy and robustness of signal analysis can be significantly improved.

[0021] Step S103, quantifying and accumulating fault features of multiple cyclic frequencies in the enhanced envelope spectrum to obtain a diagnostic feature spectrum, and identifying an equi-frequency-interval harmonic structure in the enhanced envelope spectrum according to the diagnostic feature spectrum to obtain cyclic frequencies.

[0022] In this step, the diagnostic features of each cycle frequency in the enhanced envelope spectrum are calculated according to the amplitude of the enhanced envelope spectrum, and the expression is: , In the formula, For diagnostic features, To enhance the amplitude of the envelope spectrum, is the frequency axis vector, is the background noise band considered, is the frequency; The diagnostic characteristic spectrum is constructed according to the diagnostic characteristics of each cycle frequency, and the expression is: , In the formula, For diagnostic characteristics, is the first cycle frequency, is the second cycle frequency, is the frequency of the nth cycle, is the amplitude of the n-fold enhanced envelope spectrum, is the harmonic amplitude, is the number of frequency cycles.

[0023] For a certain fundamental frequency of the harmonic structure with equal frequency intervals ,when hour, Only affected by noise, making it very small, when and When overlapped, middle The harmonic amplitude at A sharp component will appear at the point where the predicted periodic frequency may be incorrect under strong noise interference, which will eventually affect the accuracy of ACYCBD. The amplitude of indicates the significant level of non-interval harmonics in the range above the background noise level. The diagnostic feature spectrum quantifies the relatively prominent features at multiple frequencies and accumulates them.

[0024] It should be noted that the expression for calculating the cycle frequency is: , In the formula, is the cycle frequency, is the first cycle frequency, is the first cycle frequency, is the Hth cycle frequency, is the cycle frequency of the nth period, is the frequency, is the harmonic amplitude, is the number of frequency cycles.

[0025] Step S104, feeding the cyclic frequency into a preset maximum cyclostationary blind deconvolution, the maximum cyclostationary blind deconvolution outputs a repetitive transient pulse caused by a local defect, and determining a fault signal of a railway switch according to the repetitive transient pulse.

[0026] In this step, the filter is constructed, and the maximum cyclostationary blind deconvolution uses ICS2 as the objective function to determine the optimal filter, and seeks maximization through an iterative eigenvalue decomposition algorithm. In order to determine the ideal inverse filter , the evaluation index can be defined as: , In fact, the basic principle of all blind convolution algorithms is to mitigate the impact of transmission and Fault pulses are retrieved from is an ideal inverse filter, is the objective function, is an inverse FIR filter, as shown below: , in, is the target signal, For signal, is the unknown impulse response, Fault pulse, Represents the convolution operator. The matrix expression of this convolution operation is: , Among them, the cycle frequency can be described as: , in, is the cyclic frequency of the discrete-time signal, is the pulse correlation period, is the sample index, which is used as an index to characterize the periodicity of signal energy. The definition of can be expressed as: , , , , , , , in, is an indicator to characterize the periodicity of signal energy. represents the period-dependent component, is the sample index, is the sample period, is a k-order sample, is an o-order sample, is the conjugate transpose operation of the matrix, is the correlation matrix, is the sample attenuation coefficient, is the length of the signal, is the sample norm, is the number of samples, is the sample weight; Signals covering all periodic frequencies of interest Described as: , but The expression is: , Calculate the weight matrix and obtain the maximum eigenvalue , find the ideal filter , the weighted matrix Defined as: , in, and are weighted correlation matrix and correlation matrix respectively. The problem can be transformed into finding the maximum eigenvalue : , From the above formula, we can get the maximum eigenvalue and its corresponding eigenvector, maximum eigenvalue Corresponding to the best , the eigenvector is used as the ideal filter of CYCBD , Determine whether the convergence criterion is met, obtain the final filtered signal y, obtain the envelope spectrum of signal y, analyze the envelope spectrum, and complete fault diagnosis.

[0027] In summary, the method of the present application uses a signal centralized monitoring system to collect the three-phase operating current of the switch through a current sensor; then, the cyclic frequency information of the current digital signal is extracted and an enhanced envelope spectrum (EES) is generated; secondly, a diagnostic feature spectrum (DFS) is obtained by quantifying and accumulating the relative significance of multiple frequencies, which can identify the equi-frequency-interval harmonic structure (EIHS) in the EES and provide an accurate cyclic frequency; finally, the estimated cyclic frequency is fed into a maximum cyclostationary blind deconvolution (CYCBD) to retrieve repetitive transient pulses caused by local defects, thereby realizing adaptive diagnosis of railway switch faults, improving the accuracy of railway switch fault identification, and enabling fault-related signals to be well identified even in the presence of strong external noise interference, which is beneficial to the troubleshooting of railway switch faults.

[0028] See also Figure 2 , which shows a structural block diagram of a railway switch machine fault analysis system of the present application.

[0029] like Figure 2 As shown, the railway switch machine fault analysis system 200 includes an acquisition module 210 , a generation module 220 , an identification module 230 and an analysis module 240 .

[0030] The acquisition module 210 is configured to acquire the three-phase action current of the railway switch machine and convert the three-phase action current into a three-phase action current digital signal according to a preset digital conversion strategy; A generating module 220, configured to extract cyclostationarity information from the three-phase action current digital signal and generate an enhanced envelope spectrum; The identification module 230 is configured to quantify and accumulate the fault characteristics of multiple cyclic frequencies in the enhanced envelope spectrum to obtain a diagnostic characteristic spectrum, and identify the equal-frequency-interval harmonic structure in the enhanced envelope spectrum according to the diagnostic characteristic spectrum to obtain the cyclic frequency; The analysis module 240 is configured to feed the cyclic frequency into a preset maximum cyclostationary blind deconvolution, the maximum cyclostationary blind deconvolution outputs a repetitive transient pulse caused by a local defect, and determines a fault signal of a railway switch according to the repetitive transient pulse.

[0031] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A railway switch machine fault analysis method, characterized in that: include: Acquire the three-phase action current of the railway switch machine, and convert the three-phase action current into a three-phase action current digital signal according to a preset digital conversion strategy; Extracting the cycle frequency information in the three-phase action current digital signal and generating an enhanced envelope spectrum; quantifying and accumulating fault features of multiple cyclic frequencies in the enhanced envelope spectrum to obtain a diagnostic characteristic spectrum, and identifying an equi-frequency-interval harmonic structure in the enhanced envelope spectrum according to the diagnostic characteristic spectrum to obtain a cyclic frequency; The cyclic frequency is fed into a preset maximum cyclostationary blind deconvolution, the maximum cyclostationary blind deconvolution outputs a repetitive transient pulse caused by a local defect, and a fault signal of a railway switch is determined according to the repetitive transient pulse.

2. A railway switch machine fault analysis method according to claim 1, characterized in that: The digital conversion strategy is expressed as: ， ， ， ， In the formula, is the digital conversion function of the characteristic factor of the lightweight current signal, is the partial discharge signal amplitude coefficient, is a partial discharge signal with unit amplitude, is a pure partial discharge signal, For delay, To detect signal noise, is the acquisition length, is the reference signal, is the effective value of the electrical signal, is the peak value of the electrical signal, is the electrical signal pulse width, Signal duty cycle, is the main frequency component of the signal, is the signal harmonic distribution, is the signal spectrum energy entropy, For short-time overcurrent signal, For glitch signal, is the transient waveform distortion, is the signal activity, To reduce the sampling rate when the signal is stable, To trigger a high sampling rate when a mutation occurs. is the similarity of adjacent sampling points, is the effective value error, For high temperature drift, For low temperature drift, is the environmental noise, It is a weak current signal. is the adaptive threshold, is the differential noise, For noise suppression, is the noise suppression weight, is the signal time information, is the partial discharge signal weight, is the electrical signal weight.

3. A railway switch machine fault analysis method according to claim 1, characterized in that: The extracting the cycle frequency information in the three-phase action current digital signal and generating an enhanced envelope spectrum comprises: Performing short-time Fourier transformation on the three-phase action current digital signal; The transformed three-phase action current digital signal is denoised according to a preset denoising processing strategy to obtain a target three-phase action current digital signal; A spectrum correlation analysis is performed on the target three-phase action current digital signal to obtain an enhanced envelope spectrum.

4. A railway switch machine fault analysis method according to claim 3, characterized in that: The expression of the noise reduction processing strategy is: ， ， ， ， In the formula, For a stable target signal, is a non-stationary target signal, is additive noise, is a multi-component FM signal, is a non-convex regularization term, For time shift, is the L0 norm, is the signal sparsity, For time shift, is the frequency shift, To cover the signal, To tune the frequency space, is the quantity of the component, For the frequency modulation rule, is the dictionary redundancy, is the variable component, is super-Gaussian sparsity, To over-penalize large coefficients, is iterative reweighting, is the Lp norm, is the hard threshold, is the half threshold, is the pre-denoising value, To separate high frequency noise, is a low frequency signal, is the wavelet threshold, is the residual noise, To suppress the amount of impulse noise, is the noise standard deviation, is the number of dictionary atoms, are atomic parameters, is the proximal operator, is the initialization coefficient, is the time-frequency distribution, is additive Gaussian noise, As the impact component, is a transient non-stationary signal, For the wave group, For the ST segment, is the signal-to-noise ratio of the characteristic component, is a coarse-grained dictionary, is the compression coefficient vector, To incrementally update the weights, is the online calculation coefficient, For non-convex optimization of local minima, For high computational complexity, is the noise sensitivity, Automatically search for hyperparameters for Bayesian optimization, is a non-stationary target echo, is a high frequency transient signal, For high-precision time-frequency representation, is the optimization coefficient, It is a low-precision time-frequency representation.

5. A railway switch machine fault analysis method according to claim 1, characterized in that: The quantifying and accumulating the fault characteristics of multiple cycle frequencies in the enhanced envelope spectrum to obtain a diagnostic characteristic spectrum includes: According to the amplitude of the enhanced envelope spectrum, the diagnostic characteristics of each cycle frequency in the enhanced envelope spectrum are calculated, and the expression is: ， In the formula, For diagnostic features, To enhance the amplitude of the envelope spectrum, is the frequency axis vector, is the background noise band considered, is the frequency; The diagnostic characteristic spectrum is constructed according to the diagnostic characteristics of each cycle frequency, and the expression is: ， In the formula, For diagnostic characteristics, is the first cycle frequency, is the second cycle frequency, is the frequency of the nth cycle, is the amplitude of the n-fold enhanced envelope spectrum, is the harmonic amplitude, is the number of frequency cycles.

6. A railway switch machine fault analysis method according to claim 1, characterized in that: The expression for calculating the cycle frequency is: ， In the formula, is the cycle frequency, is the first cycle frequency, is the first cycle frequency, is the Hth cycle frequency, is the cycle frequency of the nth period, is the frequency, is the harmonic amplitude, is the number of frequency cycles.

7. A railway switch machine fault analysis system, characterized in that: include: An acquisition module is configured to acquire the three-phase action current of the railway switch machine and convert the three-phase action current into a three-phase action current digital signal according to a preset digital conversion strategy; A generating module configured to extract cyclostationarity information from the three-phase action current digital signal and generate an enhanced envelope spectrum; an identification module configured to quantify and accumulate fault features of multiple cyclic frequencies in the enhanced envelope spectrum to obtain a diagnostic feature spectrum, and identify an equi-frequency-interval harmonic structure in the enhanced envelope spectrum according to the diagnostic feature spectrum to obtain a cyclic frequency; The analysis module is configured to feed the cyclic frequency into a preset maximum cyclostationary blind deconvolution, the maximum cyclostationary blind deconvolution outputs a repetitive transient pulse caused by a local defect, and determines a fault signal of a railway switch according to the repetitive transient pulse.

8. An electronic device, characterized in that: include: At least one processor, and a memory communicatively connected to the at least one processor, wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the method described in any one of claims 1 to 6.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the method according to any one of claims 1 to 6 is implemented.

Images