Rack train drive gear fault diagnosis method, system and device

Abstract

The present invention relates to the field of gear fault diagnosis and is a method, system and device for diagnosing faults of drive gears of a rack train, which solves the problem in the prior art of lacking effective extraction of fault features from the vibration signal of the rack segment. The present invention comprises the following steps: identifying the vibration signal of the transition segment of the transition device; identifying the vibration signal of the rack segment; processing the vibration signal of the rack segment; and realizing fault diagnosis based on the meshing frequency and sideband information. The present invention can accurately identify the vibration signal generated by the drive gear passing through the rack segment from the vibration signal based on the time domain characteristics of the vibration signal, thereby realizing adaptive identification of the vibration signal of the rack segment. By removing the noise interference in the time-frequency distribution of the rack segment vibration signal and improving the time-frequency aggregation of the time-frequency distribution, the fault features of the drive gear are extracted from the time-frequency distribution, and the drive gear fault diagnosis is realized, so as to realize the health status monitoring of the drive gear of the rack train and ensure the safety of the train operation.

Description

Technical Field

[0001] The present invention relates to the field of gear fault diagnosis, and in particular to a method, system and device for diagnosing faults of drive gears of a rack train. Background Art

[0002] A rack train is a specialized rail vehicle designed specifically for mountainous areas. Its key difference from traditional trains lies in the addition of a rack track in the middle of the track and a drive gear at the bottom of the train. The meshing of the drive gear and rack track provides additional traction. A rack track primarily consists of a rack section, a transition section, and a wheel-rail section, with the drive gear meshing only in the rack section. Due to the extremely harsh operating environment of the drive gear, faults such as tooth flank wear and tooth root fracture are prone to occur. Failure to promptly detect these faults can not only exacerbate damage to the drive gear and rack track, reducing train reliability and passenger comfort, but can also lead to serious safety accidents.

[0003] The sideband energy ratio method for gear mesh fault detection with application number CN201110296586.7 discloses a method for evaluating degradation in a gearbox, obtaining data on the harmonic frequency amplitudes and their associated sidebands generated by the operation of the gearbox from an operating gearbox, and calculating the sideband energy ratio by dividing the amplitude of the total sideband signal associated with the harmonic by the amplitude of the harmonic; the calculation is based on the sideband amplitudes associated with adjacent harmonics, and the ratio is monitored over time, or compared with one or more values to provide an indication of degradation.

[0004] However, two key technologies are required to diagnose drive gear faults: first, accurately identifying the rack segment vibration signal from the vibration signal, and then accurately extracting the drive gear fault signature from the rack segment vibration signal. Currently, no effective method for accurately identifying rack segment vibration signals from rack train vibration signals has been established in the field of gear health monitoring. Furthermore, existing methods for extracting fault signatures lack sufficient noise suppression capabilities, making it difficult to obtain time-frequency distributions with high time-frequency aggregation, which directly affects the accuracy of fault signature extraction.

[0005] There is an urgent need for a new drive gear fault diagnosis method, system and device that can solve the above problems. Summary of the Invention

[0006] The present invention proposes a method, system and device for diagnosing rack train drive gear faults, which solve two key problems existing in the prior art: first, it is difficult to identify the rack segment vibration signal from the rack train vibration signal; second, the fault feature extraction process is seriously interfered by noise and the time-frequency resolution of the time-frequency distribution is low.

[0007] The technical solution of the present invention is achieved as follows:

[0008] A method for diagnosing a fault of a driving gear of a rack train comprises the following steps:

[0009] S1: Identify the transition section vibration signal of the transition device: Identify the periodic high-amplitude transient impact signal generated by the drive gear passing through the transition device from the vibration signal; construct an adaptive amplitude threshold , select candidate points whose amplitude exceeds the threshold from the vibration signal; based on the roller center distance of the transition device and train line speed , calculate the theoretical impact period and verify the periodicity of the candidate points; mark the data points that meet both the amplitude threshold conditions and the periodic characteristics as transition section vibration signals;

[0010] S2: Identify the rack segment vibration signal: A signal segment is intercepted from both sides of the transition segment vibration signal in S1 as a sample. The samples are quantitatively analyzed using the defined data energy mean characterization parameter. The energy mean values of the two samples are compared, and the sample with the larger data energy mean is marked as the rack segment vibration signal. The rack segment and the wheel-rail segment are located on both sides of the transition segment. When the rack train runs on the rack segment, the drive gear meshes with the rack rail. Therefore, the impact on the drive gear in the rack segment is significantly stronger than the impact on the wheel-rail segment. The rack segment vibration signal can be identified based on the time domain characteristics of the rack segment and wheel-rail segment vibration signals.

[0011] S3: Processing the rack segment vibration signal: Calculate the time-frequency distribution of the rack segment vibration signal and perform noise reduction on the time-frequency distribution to improve the time-frequency aggregation of the time-frequency distribution; Perform time-frequency analysis on the rack segment vibration signal to obtain the time-frequency distribution, divide the time-frequency distribution into multiple time slices, detect local maximum points in each time slice, and divide the peak energy divergence area with the maximum point as the center; Utilize the difference in correlation between signal components and noise to calculate the correlation between peak energy divergence areas at adjacent moments, and accordingly screen out strongly correlated areas; Strongly correlated areas are considered to be valid signal components containing fault characteristics, while the remaining areas are considered to be noise and deleted; Perform iterative processing along the time direction to obtain the time-frequency distribution with noise removed;

[0012] S4: Fault diagnosis is achieved based on the meshing frequency and sideband information: the time-frequency ridge extraction algorithm is used to extract the drive gear meshing frequency ridge and its sideband from the time-frequency distribution; based on the drive gear parameters, it is determined whether the interval between the meshing frequency and the sideband is the rotation frequency to achieve fault diagnosis.

[0013] According to a further technical solution, step S1 specifically includes the following steps:

[0014] S11 for vibration signals , set an adaptive amplitude threshold , used to accurately identify high-amplitude transient shocks in vibration signals; the threshold The calculation formula is:

[0015] (1);

[0016] in, Indicates the maximum amplitude in the vibration signal, Indicates the 0.75 quantile of the vibration signal amplitude;

[0017] From the vibration signal Filtering exceeds threshold High-amplitude transient shock:

[0018] (2);

[0019] in, represents the amplitude that satisfies the threshold condition, Indicates the corresponding time, Indicates that the point has no value;

[0020] S12 Since the driving gear usually passes through the transition section at a uniform speed, the impact generated by the collision between the driving gear and the roller of the transition device has a significant periodic characteristic; based on the distance between the center of the roller of the transition device and train line speed , calculate the theoretical impact period, and use it to analyze the time series Perform periodicity test, and the expression of periodicity at the test time point is:

[0021] (3);

[0022] in, Indicates the time value that meets the impact cycle condition, Indicates the center distance between the two rollers in the transition device, represents the linear speed of the train, Indicates that the point has no value;

[0023] Since the impact period is constrained in the interval Therefore, there is naturally periodicity between the time points that meet the constraint conditions; the time points that meet the periodicity judgment conditions The corresponding data points are marked as transition section vibration signals.

[0024] Further technical solution, the step S2 is specifically:

[0025] The rack section S21 and the wheel-rail section S21 are located on both sides of the transition section. When the rack train runs in the rack section, the driving gear meshes with the rack rail. Therefore, the impact on the driving gear in the rack section is significantly stronger than that in the wheel-rail section. By analyzing the time domain characteristics of the vibration signals of the rack section and the wheel-rail section, the rack section vibration signal can be identified. A section of the signal on both sides of the transition section vibration signal output by S12 is intercepted as a sample and set as and ;

[0026] Next, define a data energy mean characterization parameter:

[0027] (4);

[0028] in, Indicates the maximum amplitude in the vibration signal, Represents the number of sample data points intercepted, and this formula is used to quantify the mean data energy; represents the mean energy;

[0029] When the rack train is running on the rack section, the impact on the driving gear is significantly stronger than the impact on the driving gear on the wheel rail due to the meshing of the driving gear and the rack. Based on this characteristic, the energy mean values calculated from the two selected samples are compared, and the data with the larger energy mean value is defined as the rack section vibration signal:

[0030] Rack segment vibration signal = (5);

[0031] in, Representation data The calculated mean energy is Representation data Calculate the mean energy.

[0032] Further technical solution, the step S3 is specifically:

[0033] S31 divides the peak energy divergence area: performs short-time Fourier transform on the rack vibration signal output by S22 to obtain the time-frequency distribution of the rack vibration signal , and denoise the time-frequency distribution; based on Search for the amplitude of the data column along the frequency axis The local maximum set of and local minimum sets :

[0034] (6);

[0035] (7);

[0036] in, represents a local maximum, represents a local minimum, Indicates The moment frequency value is The amplitude of Indicates The moment frequency value is The amplitude of is the frequency step;

[0037] For each local maximum , search for the local minimum values adjacent to it and , forming a peak energy divergence area :

[0038] (8);

[0039] but The set of all peak energy divergence areas divided by time Defined as:

[0040] (9);

[0041] in, for The number of local maxima at time instant;

[0042] S32 screening of strong correlation areas in the peak energy divergence region: select At some point Start the analysis, assuming All peak energy divergence areas in are signal components; then, Select a peak energy divergence area , and in Search for a peak energy divergence area closest to it along the frequency axis , calculate the Pearson correlation coefficient between the two , and repeat the above steps until All peak energy divergence areas at the moment are calculated once;

[0043] (10);

[0044] (11);

[0045] in, express time TFR The mean amplitude of represents the Pearson correlation coefficient result, express Moment Peak energy divergence area; Indicates the s discrete frequency points; Indicates the number of discrete points in the frequency direction; represents a frequency variable;

[0046] If the correlation coefficient result is greater than or equal to 0.7, it is determined to be a strong correlation and retained. The peak energy divergence area corresponding to the moment; otherwise, the area is considered as noise and is directly deleted The peak energy divergence area corresponding to the moment;

[0047] (12);

[0048] in, express The set of peak energy divergence areas that are retained at all times, Indicates that the point has no value;

[0049] S33 removes the noise of time-frequency distribution: along the time direction, by Repeat the S32 method at both ends until the peak energy divergence area of all moments is found, and a time-frequency distribution with noise removed is output. .

[0050] As a further technical solution, the step S3 further includes S34 to improve the time-frequency aggregation of the time-frequency distribution: using a time reordering operator to compensate for the theoretical estimation deviation of the original frequency reordering operator, and proposing a frequency compensation reordering operator:

[0051] (13);

[0052] in, represents the time derivative, represents the instantaneous amplitude, represents the instantaneous phase, represents the instantaneous frequency, represents the window function, represents an imaginary number; represents the time-frequency distribution after removing noise;

[0053] According to the frequency compensation rearrangement operator The estimated instantaneous frequency redistributes the divergent time-frequency coefficients to the energy trajectory along the frequency direction, further improving the time-frequency aggregation of the time-frequency distribution;

[0054] (14);

[0055] in, is the time-frequency distribution of the frequency-compensated synchrocompression transform, represents the Dirac function.

[0056] Further technical solution, the step S4 is specifically: using the ridge extraction algorithm, from Extract the meshing frequency of the drive gear and rack and sidebands, according to the gear parameters, Convert it into the rotational frequency of the driving gear, determine whether the interval between the meshing frequency and the sideband is the rotational frequency, and realize fault diagnosis.

[0057] The rack train drive gear fault diagnosis system includes a data acquisition module, a rack segment data identification module, a rack segment data processing module and a drive gear fault diagnosis module connected in sequence; the data acquisition module includes a vibration sensor installed near the rack train drive gear; used to collect vibration signals; the data acquisition module is responsible for converting the analog signals collected by the sensor into digital signals and transmitting them to the data processing unit; the data processing unit runs the corresponding algorithm program; wherein the rack segment data identification module marks samples with large data energy mean as rack segment vibration signals and outputs the rack segment vibration signals; the rack segment data processing module processes the rack segment vibration signals and outputs a frequency-compensated synchronous compression transform time-frequency distribution The drive gear fault diagnosis module uses the time-frequency ridge extraction algorithm to extract the time-frequency distribution The driving gear meshing frequency ridge and its sidebands are extracted from the meshing frequency of the driving gear to analyze the type and severity of the fault.

[0058] The present invention also discloses a rack train drive gear fault diagnosis device, which is provided with the above-mentioned rack train drive gear fault diagnosis system and also includes a display device for displaying the fault diagnosis results, a power module; a communication module for transmitting the vibration signal of the vibration sensor; the power module supplies power to the communication module, the display device and the fault diagnosis system.

[0059] Compared with the prior art, the rack train drive gear fault diagnosis method and system disclosed in the present invention have the following beneficial effects.

[0060] 1. According to the time domain characteristics of the vibration signal, the vibration signal generated by the driving gear passing through the rack segment can be accurately identified from the vibration signal, thereby realizing adaptive recognition of the rack segment vibration signal.

[0061] 2. By removing noise interference from the time-frequency distribution of the rack segment vibration signal and improving the time-frequency aggregation of the time-frequency distribution, the fault characteristics of the drive gear are extracted from the time-frequency distribution, and drive gear fault diagnosis is realized. This allows the health status of the rack train drive gear to be monitored and the safe operation of the train to be ensured. BRIEF DESCRIPTION OF THE DRAWINGS

[0062] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0063] Figure 1 : Schematic diagram of this application

[0064] Figure 2 : Schematic diagram of the rack track of this application;

[0065] Figure 3 : Flow chart of the rack segment data identification module in this application.

[0066] Figure 4 : Flow chart of the rack segment data processing module in this application.

[0067] Figure 5 : Schematic diagram of the flow chart of the drive gear fault diagnosis module in this application. DETAILED DESCRIPTION

[0068] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention without making any creative efforts shall fall within the scope of protection of the present invention.

[0069] In this embodiment, the vibration sensor, data acquisition module, data processing unit and corresponding data processing software installed on the rack train are used to achieve this. Figure 1 As shown in the system schematic diagram of a rack train drive gear fault diagnosis method of the present application, a vibration sensor is installed near the rack train drive gear to collect vibration signals; the data acquisition module is responsible for converting the analog signal collected by the sensor into a digital signal and transmitting it to the data processing unit; the data processing unit runs the corresponding algorithm program, and the data processing unit includes a rack segment data recognition module, a rack segment data processing module and a drive gear fault diagnosis module; the collected vibration signal is analyzed and processed to realize the fault diagnosis of the drive gear.

[0070] like Figure 2 The rack track diagram of the rack train driving gear fault diagnosis method of the present application is shown as follows: the rack train runs on a specific rack railway line, which includes a wheel-rail section, a transition section and a rack track section. The linear speed of the train during operation has been set to 0.11 meters through precise measurement. The train's speed sensor can be used to obtain data in real time. Meanwhile, a high-precision vibration sensor is installed near the drive gear of the rack train. This sensor can collect vibration signals during the train's operation at a sampling frequency of 10kHz. The collected vibration signals are then transmitted from the data acquisition module to the data processing unit for subsequent analysis.

[0071] like Figure 3 The process diagram of the rack segment data identification module in this application is shown as follows: After the data processing unit receives the vibration signal collected by the vibration sensor, it constructs an adaptive amplitude threshold Specifically, according to the preset algorithm, the amplitude of the vibration signal is statistically analyzed to obtain the adaptive amplitude threshold Based on the threshold, the points with amplitudes exceeding the threshold are selected from the vibration signal as candidate points. and train line speed , calculate the theoretical shock cycle The theoretical impact cycle represents the time interval between two adjacent impacts when the driving gear passes through the transition device. Periodic verification is performed on the selected candidate points, that is, checking whether the time interval between adjacent candidate points is consistent with the theoretical impact cycle. If the error is within a certain range (±5%), it is considered to meet the periodicity. The data points with conditional and periodic characteristics are marked as transition section vibration signals.

[0072] From both sides of the transition vibration signal marked in the above step, a 10-second signal with a corresponding signal data volume of 100k is intercepted as a sample. Define the data energy mean characterization parameter , which is calculated by summing the amplitude of each sample point in the sample signal, multiplying it by the maximum amplitude in the sample signal, and then dividing it by the number of sample points. This parameter is used to quantitatively analyze two samples and compare their mean energy values. Because the drive gear meshes with the rack rail when a rack train is operating in the rack section, the impact it experiences is significantly stronger than that in the wheel-rail section. Therefore, samples with large mean energy values are marked as rack section vibration signals and output as rack section vibration signals.

[0073] like Figure 4 As shown, the rack segment vibration signal is processed: the time-frequency distribution of the rack segment vibration signal is calculated using the short-time Fourier transform method. . In order to remove the interference of noise in the time-frequency distribution, a noise reduction algorithm is used for the time-frequency distribution. The time-frequency distribution is divided into multiple time slices. In each time slice, the local maximum point and the local minimum point are detected by the local search algorithm. The peak energy divergence area is divided with the detected maximum point as the center and the two adjacent minimum points as the boundary. The correlation degree between the peak energy divergence areas at adjacent moments can be calculated by using the difference in correlation between the signal component and the noise. Specifically, it can be judged by calculating the correlation coefficient, etc., and a correlation coefficient threshold is set to 0.7. The area with a correlation degree greater than the threshold is regarded as a strong correlation area. The strong correlation area is regarded as a valid signal component containing fault characteristics, and the remaining areas are regarded as noise and deleted. Iterative processing is performed along the time direction, and the above operations of dividing time slices, detecting maximum points, detecting adjacent minimum points, dividing peak energy divergence areas, screening strong correlation areas, etc. are repeated until the time-frequency distribution of the entire rack segment vibration signal is processed to obtain the time-frequency distribution with noise removed. ; Obtain more accurate instantaneous frequency estimates through frequency compensation reordering operators; Redistribute the time-frequency coefficients of the time-frequency distribution , output frequency compensation synchronous compression transform time-frequency distribution .

[0074] like Figure 5 As shown, the time-frequency ridge extraction algorithm is used to extract the time-frequency distribution Extract the driving gear meshing frequency ridge and its sideband. According to the number of teeth of the driving gear , converting the meshing frequency into the rotational frequency of the drive gear. Determine whether the interval between the meshing frequency and the sideband is the rotational frequency or an integer multiple of the rotational frequency. If it meets this rule, it is determined that the drive gear has a fault, such as tooth surface wear, tooth root fracture, etc., and the fault diagnosis result is output; if it does not meet the rule, it is preliminarily determined that the drive gear is operating normally. At the same time, by analyzing the distribution characteristics of the meshing frequency and its sidebands (such as the number of sidebands, amplitude change pattern and distribution range), the fault type can be accurately identified: if the number of sidebands on both sides of the meshing frequency is small but the amplitude fluctuates violently and the distribution is concentrated, it is characterized by tooth surface wear fault; if a large number of sidebands with low amplitude, uniform distribution and flatness appear, it is characterized by tooth root fracture fault.

[0075] Through the above steps, the fault diagnosis of the drive gear of the rack railway train is realized, and potential fault problems of the drive gear can be discovered in time to ensure the safe operation of the rack railway train.

[0076] Specifically, the rack train drive gear fault diagnosis method includes the following steps:

[0077] S1: Identify the transition section vibration signal of the transition device: Identify the periodic high-amplitude transient impact signal generated by the drive gear passing through the transition device from the vibration signal; construct an adaptive amplitude threshold , select candidate points whose amplitude exceeds the threshold from the vibration signal; based on the roller center distance of the transition device and train line speed , calculate the theoretical impact period and verify the periodicity of the candidate points; mark the data points that meet both the amplitude threshold conditions and the periodic characteristics as transition section vibration signals;

[0078] S11 for vibration signals , set an adaptive amplitude threshold ; The threshold The calculation formula is:

[0079] (1);

[0080] in, Indicates the maximum amplitude in the vibration signal, Indicates the 0.75 quantile of the vibration signal amplitude;

[0081] From the vibration signal Filtering exceeds threshold High-amplitude transient shock:

[0082] (2);

[0083] in, represents the amplitude that satisfies the threshold condition, Indicates the corresponding time, Indicates that the point has no value;

[0084] S12 roller center distance based on transition device and train line speed , calculate the theoretical impact period, and use it to analyze the time series Perform periodicity test, and the expression of periodicity at the test time point is:

[0085] (3);

[0086] in, Indicates the time value that meets the impact cycle condition, Indicates the center distance between the two rollers in the transition device, represents the linear speed of the train, Indicates that the point has no value;

[0087] The time point that satisfies the periodicity judgment condition The corresponding data points are marked as transition section vibration signals.

[0088] S2: Identify the rack segment vibration signal: A segment of the signal is intercepted from both sides of the transition segment vibration signal in S1 as a sample. The sample is quantitatively analyzed using the defined data energy mean characterization parameter. The energy mean values of the two samples are compared, and the sample with the larger data energy mean is marked as the rack segment vibration signal.

[0089] S21 intercepts a section of the signal on both sides of the transition section vibration signal output by S12 as a sample, and sets them as and ;

[0090] Next, define a data energy mean characterization parameter:

[0091] (4);

[0092] in, Indicates the maximum amplitude in the vibration signal, Represents the number of sample data points intercepted, and this formula is used to quantify the mean data energy; represents the mean energy;

[0093] S22 compares the energy mean values calculated from the two selected samples and defines the data with the larger energy mean value as the rack segment vibration signal:

[0094] Rack segment vibration signal = (5);

[0095] in, Representation data The calculated mean energy is Representation data Calculate the mean energy.

[0096] S3: Processing the rack segment vibration signal: Perform time-frequency analysis on the rack segment vibration signal to obtain the time-frequency distribution. The time-frequency distribution is divided into multiple time slices. Local maximum points are detected within each time slice, and the peak energy divergence area is divided around the maximum point. The difference in correlation between signal components and noise is used to calculate the correlation between peak energy divergence areas at adjacent moments, and strongly correlated areas are screened accordingly. Strongly correlated areas are considered to be valid signal components containing fault characteristics. Iterative processing is performed along the time direction to obtain the time-frequency distribution with noise removed.

[0097] S31 divides the peak energy divergence area: performs short-time Fourier transform on the rack vibration signal output by S22 to obtain the time-frequency distribution of the rack vibration signal , and denoise the time-frequency distribution; based on Search for the amplitude of the data column along the frequency axis The local maximum set of and the set of local minima :

[0098] (6);

[0099] (7);

[0100] in, represents a local maximum, represents a local minimum, Indicates The moment frequency value is The amplitude of Indicates The moment frequency value is The amplitude of is the frequency step;

[0101] For each local maximum , search for the local minimum values adjacent to it and , forming a peak energy divergence area :

[0102] (8);

[0103] but The set of all peak energy divergence areas divided by time Defined as:

[0104] (9);

[0105] in, for The number of local maxima at time instant;

[0106] S32 screening peak energy divergence area with strong correlation:

[0107] choose At some point Start the analysis, assuming All peak energy divergence areas in are signal components; then, Select a peak energy divergence area , and in Search for a peak energy divergence area closest to it along the frequency axis , calculate the Pearson correlation coefficient between the two , and repeat the above steps until All peak energy divergence areas at the moment are calculated once;

[0108] (10);

[0109] (11);

[0110] in, express time TFR The mean amplitude of represents the Pearson correlation coefficient result, express Moment Peak energy divergence area; Indicates the s discrete frequency points; Indicates the number of discrete points in the frequency direction; represents a frequency variable;

[0111] If the correlation coefficient result is greater than or equal to 0.7, it is determined to be a strong correlation and retained. The peak energy divergence area corresponding to the moment; otherwise, the area is considered as noise and is directly deleted The peak energy divergence area corresponding to the moment;

[0112] (12);

[0113] in, express The set of peak energy divergence areas that are retained at all times, Indicates that the point has no value;

[0114] S33 removes the noise of time-frequency distribution: along the time direction, by Repeat the S32 method at both ends until the peak energy divergence area of all moments is found, and a time-frequency distribution with noise removed is output. .

[0115] S34 improves the time-frequency aggregation of time-frequency distribution: a frequency compensation rearrangement operator is proposed:

[0116] (13);

[0117] in, represents the time derivative, represents the instantaneous amplitude, represents the instantaneous phase, represents the instantaneous frequency, represents the window function, represents an imaginary number; represents the time-frequency distribution after noise removal;

[0118] According to the frequency compensation rearrangement operator The estimated instantaneous frequency redistributes the divergent time-frequency coefficients onto the energy trajectory along the frequency direction;

[0119] (14);

[0120] in, is the time-frequency distribution of the frequency-compensated synchrocompression transform, represents the Dirac function.

[0121] S4: Fault diagnosis is achieved based on the meshing frequency and sideband information: the time-frequency ridge extraction algorithm is used to extract the drive gear meshing frequency ridge and its sideband from the frequency-compensated synchronous compression transform time-frequency distribution; based on the drive gear parameters, it is determined whether the interval between the meshing frequency and the sideband is the rotation frequency to achieve fault diagnosis.

[0122] Through the ridge extraction algorithm, Extract the meshing frequency of the driving gear and rack and sidebands, according to the gear parameters, Convert it into the rotational frequency of the driving gear, determine whether the interval between the meshing frequency and the sideband is the rotational frequency, and realize fault diagnosis.

[0123] On the other hand, the present invention provides a rack train drive gear fault diagnosis system, comprising a data acquisition module, a rack segment data identification module, a rack segment data processing module and a drive gear fault diagnosis module connected in sequence; the data acquisition module comprises a vibration sensor installed near the rack train drive gear; used to collect vibration signals; the data acquisition module is responsible for converting the analog signal collected by the sensor into a digital signal and transmitting it to a data processing unit; the data processing unit runs a corresponding algorithm program; wherein the rack segment data identification module marks samples with a large data energy mean as rack segment vibration signals and outputs the rack segment vibration signals; the rack segment data processing module processes the rack segment vibration signals and outputs a frequency-compensated synchronous compression transform time-frequency distribution The drive gear fault diagnosis module uses the time-frequency ridge extraction algorithm to extract the time-frequency distribution The driving gear meshing frequency ridge and its sidebands are extracted from the meshing frequency of the driving gear to analyze the type and severity of the fault.

[0124] The present invention also discloses a rack train drive gear fault diagnosis device, which is provided with the above-mentioned rack train drive gear fault diagnosis system and also includes a display device for displaying fault diagnosis results, a power module; a communication module for transmitting the vibration signal of the vibration sensor; the power module supplies power to the communication module, the display device and the fault diagnosis system.

[0125] Of course, without departing from the spirit and essence of the present invention, technicians familiar with the field should be able to make various corresponding changes and modifications based on the present invention, but these corresponding changes and modifications should all fall within the scope of protection of the claims attached to the present invention.

Claims

1. A method for diagnosing a fault in a rack train driving gear, characterized by: The following steps are involved: S1: Identify the transition section vibration signal of the transition device: Identify the periodic high-amplitude transient impact signal generated by the drive gear passing through the transition device from the vibration signal; construct an adaptive amplitude threshold , select candidate points whose amplitude exceeds the threshold from the vibration signal; based on the roller center distance of the transition device and train line speed , calculate the theoretical impact period and perform periodic verification on the candidate points; The data points that meet both the amplitude threshold condition and the periodic characteristics are marked as transition section vibration signals; specifically, the following steps are included: S11 for vibration signals , set an adaptive amplitude threshold ; The threshold The calculation formula is: (1)； in, Indicates the maximum amplitude in the vibration signal, Indicates the 0.75 quantile of the vibration signal amplitude; From the vibration signal Filtering exceeds threshold High-amplitude transient shock: (2)； in, represents the amplitude that meets the threshold condition, Indicates the corresponding time, Indicates that the point has no value; S12 roller center distance based on transition device and train line speed , calculate the theoretical impact period, and use it to analyze the time series Perform periodicity test, and the expression of periodicity at the test time point is: (3)； in, Indicates the time value that meets the impact cycle condition, Indicates the center distance between the two rollers in the transition device, represents the linear speed of the train, Indicates that the point has no value; The time point that satisfies the periodicity judgment condition The corresponding data points are marked as transition section vibration signals; S2: Identify the rack segment vibration signal: A segment of the signal is intercepted from both sides of the transition segment vibration signal in S1 as a sample. The sample is quantitatively analyzed using the defined data energy mean characterization parameter. The energy mean values of the two samples are compared, and the sample with the larger data energy mean is marked as the rack segment vibration signal. S3: Processing the rack segment vibration signal: Perform time-frequency analysis on the rack segment vibration signal to obtain the time-frequency distribution. The time-frequency distribution is divided into multiple time slices. Local maximum points are detected within each time slice, and the peak energy divergence area is divided around the maximum point. The difference in correlation between signal components and noise is used to calculate the correlation between peak energy divergence areas at adjacent moments, and strongly correlated areas are screened accordingly. Strongly correlated areas are considered to be valid signal components containing fault characteristics. Iterative processing is performed along the time direction to obtain the time-frequency distribution with noise removed. S4: Fault diagnosis is achieved based on the meshing frequency and sideband information: the time-frequency ridge extraction algorithm is used to extract the drive gear meshing frequency ridge and its sideband from the frequency-compensated synchronous compression transform time-frequency distribution; based on the drive gear parameters, it is determined whether the interval between the meshing frequency and the sideband is the rotation frequency to achieve fault diagnosis.

2. The rack train driving gear fault diagnosis method according to claim 1, characterized in that: The specific steps of step S2 are: S21 intercepts a section of the signal on both sides of the transition section vibration signal output by S12 as a sample, and sets them as and ; Next, define a data energy mean characterization parameter: (4)； in, Indicates the maximum amplitude in the vibration signal, Represents the number of sample data points intercepted, and this formula is used to quantify the mean data energy; represents the mean energy; S22 compares the energy mean values calculated from the two selected samples and defines the data with the larger energy mean value as the rack segment vibration signal: Rack segment vibration signal = (5); in, Representation data The calculated mean energy is Representation data Calculate the mean energy.

3. The rack train driving gear fault diagnosis method according to claim 2, characterized in that: The specific steps of step S3 are: S31 divides the peak energy divergence area: performs short-time Fourier transform on the rack vibration signal output by S22 to obtain the time-frequency distribution of the rack vibration signal , and denoise the time-frequency distribution; based on Search for the amplitude of the data column along the frequency axis The local maximum set of and local minimum sets : (6)； (7)； in, represents the local maximum frequency value, represents the local minimum frequency value, Indicates The moment frequency value is The amplitude of Indicates The moment frequency value is The amplitude of is the frequency step; For each local maximum , search for the local minimum values adjacent to it and , forming a peak energy divergence area : (8)； but The set of all peak energy divergence areas divided by time Defined as: (9)； in, for The number of local maxima at time instant; S32 screening peak energy divergence area with strong correlation: choose At some point Start the analysis, assuming All peak energy divergence areas in are signal components; then, Select a peak energy divergence area , and in Search for a peak energy divergence area closest to it along the frequency axis , calculate the Pearson correlation coefficient between the two , and repeat the above steps until All peak energy divergence areas at the moment are calculated once; (10)； (11)； in, express time TFR The mean amplitude of represents the Pearson correlation coefficient result, express Moment Peak energy divergence area; Indicates the s discrete frequency points; Indicates the number of discrete points in the frequency direction; represents a frequency variable; If the correlation coefficient result is greater than or equal to 0.7, it is determined to be a strong correlation and retained. The peak energy divergence area corresponding to the moment; otherwise, the area is considered as noise and is directly deleted The peak energy divergence area corresponding to the moment; (12)； in, express The set of peak energy divergence areas that are retained at all times, Indicates that the point has no value; S33 removes the noise of time-frequency distribution: along the time direction, by Repeat the S32 method at both ends until the peak energy divergence area of all moments is found, and a time-frequency distribution with noise removed is output. .

4. The rack train driving gear fault diagnosis method according to claim 3, characterized in that: The step S3 further includes S34 to improve the time-frequency aggregation of the time-frequency distribution: proposing a frequency compensation rearrangement operator: (13)； in, represents the time derivative, represents the instantaneous amplitude, represents the instantaneous phase, represents the instantaneous frequency, represents the window function, represents an imaginary number; represents the time-frequency distribution after removing noise; According to the frequency compensation rearrangement operator The estimated instantaneous frequency redistributes the divergent time-frequency coefficients onto the energy trajectory along the frequency direction; (14)； in, is the time-frequency distribution of the frequency-compensated synchrocompression transform, represents the Dirac function.

5. The rack train driving gear fault diagnosis method according to claim 4, characterized in that: The step S4 specifically comprises: using a ridge extraction algorithm to extract Extract the meshing frequency of the driving gear and rack and sidebands, according to the gear parameters, Convert it into the rotational frequency of the driving gear, determine whether the interval between the meshing frequency and the sideband is the rotational frequency, and realize fault diagnosis.

6. A rack train driving gear fault diagnosis system, using the rack train driving gear fault diagnosis method according to any one of claims 1 to 5, characterized in that: It includes a data acquisition module, a rack segment data identification module, a rack segment data processing module and a driving gear fault diagnosis module connected in sequence; The data acquisition module includes a vibration sensor installed near the driving gear of the rack train; it is used to collect vibration signals; the data acquisition module is responsible for converting the analog signal collected by the sensor into a digital signal and transmitting it to the data processing unit; the data processing unit runs the corresponding algorithm program; The rack segment data identification module marks samples with large data energy mean as rack segment vibration signals and outputs the rack segment vibration signals; The rack segment data processing module processes the rack segment vibration signal and outputs frequency compensation synchronous compression transform time-frequency distribution ; The driving gear fault diagnosis module uses the time-frequency ridge extraction algorithm to extract the time-frequency distribution The driving gear meshing frequency ridge and its sidebands are extracted from the meshing frequency of the driving gear to analyze the type and severity of the fault.

7. A rack train driving gear fault diagnosis device, comprising the rack train driving gear fault diagnosis system according to claim 6, characterized in that: It also includes a display device for displaying fault diagnosis results, a power module; a communication module for transmitting the vibration signal of the vibration sensor; the power module supplies power to the communication module, the display device and the fault diagnosis system.

Images