Train vibration noise feature extraction method for rail corrugation identification

Abstract

The invention provides a rail corrugation identification-oriented train vibration noise feature extraction method, which comprises the following steps of S1, determining cut-off frequencies of noise and vibration in a train according to main distribution features of line corrugation wavelengths and a train running speed; s2, band-pass filtering is carried out on noise and vibration signals in the vehicle based on the cut-off frequency; s3, carrying out wave crest and wave trough identification on the vibration noise signal after band-pass filtering, and calculating a wave depth index, namely vibration and noise characteristics, and S4, taking the distance between wave crests as the rail corrugation wavelength, in practical application, the vibration noise response in a carriage can be acquired through a lightweight sensing technology, the data acquisition cost is relatively low, the data acquisition mode is more convenient and rapid, and the method is suitable for popularization and application. And the rapid measurement requirement of subway line space-time dense steel rail corrugation can be met.

Description

Technical Field

[0001] The present invention relates to a method for extracting vibration and noise (i.e., vibration acceleration and noise) features in a railway train, and in particular to a method for extracting train vibration and noise features for rail corrugation identification, which can be applied to the rapid identification of rail corrugation and improve the detection efficiency of railway engineering departments. Background Art

[0002] Rail corrugation is a periodic, wave-shaped uneven wear on the rail surface. It is the main cause of severe high-frequency vibration in the wheel-rail system and is more common in subway lines. The environmental vibration induced by rail corrugation is one of the seven major environmental hazards in the world, which will seriously affect the comfort of subway rides, the normal life of residents along the subway line, and the normal use of precision equipment.

[0003] In addition, under the long-term effect of rail corrugation, key components of the track and vehicle will suffer fatigue damage, such as fastener bouncing and breaking, axial cracks, etc., which seriously threaten the safety of subway operation. Therefore, achieving efficient identification of the corrugation status of subway rails has important engineering value for guiding the subway engineering department to formulate grinding operation plans and ensure the safety and comfort of the subway.

[0004] At present, the most common rail corrugation detection methods can be divided into four categories. The first method mainly uses a fixed-length ruler to detect corrugation. In this type of detection method, maintenance personnel first used feeler gauges to determine the state of rail corrugation. During the detection, the wave depth of rail corrugation is estimated by the number of steel sheets that can be inserted. Usually, the feeler gauge cannot directly measure the wavelength and range of rail corrugation. Therefore, railway maintenance personnel use rail flat rulers to measure the wave depth and wavelength of rail corrugation at the same time. Considering that the reading of the rail flat ruler is very cumbersome, electronic straightness meters are widely used to measure the service status of rail corrugation. However, the fixed-length ruler can only measure the rail corrugation status at a fixed position. In order to measure rail corrugation over a large range, such equipment needs to be frequently moved.

[0005] In view of this, rail corrugation trolleys (Category 2) are also commonly used for rail corrugation detection. Most rail corrugation trolleys use high-performance sensors with high measurement accuracy, but they are expensive and difficult to achieve large-scale application. At the same time, although the detection efficiency of rail corrugation trolleys (such as CAT, BBM, etc.) is improved compared to fixed-length rulers, their measurement speed is mostly less than 5km / h. Generally, they can only detect 1-2 subway sections within a skylight point, making it difficult to quickly identify the corrugation status of the entire line.

[0006] Therefore, many scholars and researchers install sensor equipment on the train axle box or bogie to measure and evaluate the state of rail corrugation (Category 3). However, this type of detection method requires large-scale track detection equipment, which has high cost and use cost. In addition, installing sensors on the bogie and train axle box requires applying for authorization from the vehicle management department, which is a very cumbersome process.

[0007] In recent years, considering that it is relatively easy to obtain data such as vibration and noise in the car, lightweight and convenient intelligent vehicle-mounted terminals (Category 4) have gradually been adopted as a solution for rapid detection of rail corrugation. Many studies have proved that the vibration and noise response in the car can reflect the true state of rail corrugation, and compared with the other three types of detection technologies, the vibration and noise response in the car can be collected through lightweight sensing technology, the data acquisition cost is lower, and the data collection method is more convenient and faster, which can meet the needs of rapid measurement of rail corrugation on subway lines with dense time and space (the entire subway line can be inspected within 2-3 hours). Summary of the invention

[0008] The problems existing in the prior art solved by this application are: The existing technology lacks a deep mining method for train in-car vibration and noise data to identify rail corrugation. Therefore, a new method is needed to propose in-car vibration and noise data features to identify rail corrugation.

[0009] The present invention solves the technical problem by providing a train noise feature extraction method for rail corrugation identification, comprising the following steps: S1: Determine the cutoff frequency of noise and vibration inside the train based on the main distribution characteristics of the line corrugation wavelength and the train running speed; S2: Bandpass filtering of the noise and vibration signals inside the vehicle based on the cutoff frequency; S3: Identify the peaks and valleys of the vibration and noise signal after bandpass filtering, and calculate the wave depth index, i.e., the vibration and noise characteristics; S4: The distance between wave crests is taken as the rail corrugation wavelength.

[0010] Preferably, in step S1, there is the following relationship between the main vibration frequency of the in-vehicle vibration and noise signal and the rail corrugation wavelength: ; In the formula, is the train speed (m / s); is the wavelength of rail corrugation (m), which can be obtained based on existing statistical data; is the main frequency of the noise and vibration signal inside the car (Hz).

[0011] Preferably, the train speed and rail corrugation wavelength vary, and with other noise interference inside the train, the noise and vibration signals inside the train include upper and lower cutoff frequencies, namely: ; In the formula, and are the upper and lower cut-off frequencies of the bandpass filter respectively.

[0012] Preferably, in step S2, the timing noise signal (denoted as ) to perform filtering processing to obtain the characteristics of the noise signal in the corresponding frequency band: ; In the formula, is the noise signal characteristic in a specific frequency band, is the timing noise signal The frequency domain expression of is calculated by Fourier transform; ; It is a bandpass filter, and its characteristics are expressed as follows: ; In the formula, and are the upper and lower cutoff frequencies of the bandpass filter respectively; in actual situations, the upper and lower cutoff frequencies of the bandpass filter can be further determined based on the time-frequency analysis results of the noise signal; the filtered time series noise signal is recorded as , can be obtained by inverse Fourier transform: .

[0013] Preferably, in step S3, after the noise data is calibrated by mileage, the filtered time series noise signal can be Converted into spatial sequence data along the route ; For the filtered noise spatial sequence To identify the peak value, the peak value must meet the following conditions at the same time: ; In the formula, is the mileage position of the peak value, and They are exist The first and second derivatives at Defined as wave depth index.

[0014] Preferably, in step S4, after obtaining the peak position, the characteristic wavelength can be estimated according to the mileage difference at adjacent peak positions. ,in, and They are Mileage positions of adjacent crests.

[0015] The technical effects of solving the technical problem in this application are as follows: Compared with the prior art, the train noise and vibration feature extraction method for rail corrugation identification of the present invention adopts the following steps simultaneously: S1: determining the cutoff frequency of the noise and vibration in the car according to the main distribution characteristics of the line corrugation wavelength and the train running speed; S2: performing band-pass filtering on the noise and vibration signals in the car based on the cut-off frequency; S3: identifying the peaks and troughs of the noise and vibration signals after band-pass filtering, and calculating the wave depth index, that is, the vibration and noise characteristics; S4: taking the distance between the peaks as the rail corrugation wavelength. In practical applications, the noise and vibration response in the car can be collected by lightweight sensing technology, the data acquisition cost is low, and the data collection method is more convenient and quick, which can meet the needs of rapid measurement of rail corrugation with dense time and space on subway lines. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 It is a flow chart of a train noise feature extraction method for rail corrugation identification according to the present invention.

[0017] Figure 2 It is a noise signal peak recognition and a local enlarged view of the noise signal peak recognition in a train noise feature extraction method for rail corrugation identification of the present invention.

[0018] Figure 3 It is a schematic diagram of the characteristic of the in-vehicle noise signal extraction in the train noise feature extraction method for rail corrugation identification of the present invention - wave depth index.

[0019] Figure 4 The present invention is a correlation diagram between a noise wave depth index and a rail corrugation wave depth in a train noise feature extraction method for rail corrugation identification. DETAILED DESCRIPTION

[0020] In order to make the purpose, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments.

[0021] It should be understood that the specific embodiments described herein are only used to explain the present invention, and are not used to limit the present invention.

[0022] Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in the field without making any creative work shall fall within the scope of protection of the present invention.

[0023] It should be noted that when a component is said to be "fixed to" another component, it can be directly on the other component or there can be a central component. When a component is said to be "connected to" another component, it can be directly connected to the other component or there can be a central component at the same time.

[0024] Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.

[0025] The terms used in the specification of the present invention herein are only for the purpose of describing specific embodiments and are not intended to limit the present invention. The term "and / or" used herein includes any and all combinations of one or more of the related listed items.

[0026] Some embodiments of the present invention are described in detail below in conjunction with the accompanying drawings. In the absence of conflict, the following embodiments and features in the embodiments can be combined with each other.

[0027] See also Figures 1 to 4 The present invention provides a train noise feature extraction method 1 for rail corrugation identification, comprising the following steps: S1: Determine the cutoff frequency of noise and vibration inside the train based on the main distribution characteristics of the line corrugation wavelength and the train running speed; S2: Bandpass filtering of the noise and vibration signals inside the vehicle based on the cutoff frequency; S3: Identify the peaks and valleys of the vibration and noise signal after bandpass filtering, and calculate the wave depth index, i.e., the vibration and noise characteristics; S4: The distance between wave crests is taken as the rail corrugation wavelength.

[0028] The present application adopts the following steps simultaneously: S1: determining the cutoff frequency of the noise and vibration inside the car according to the main distribution characteristics of the line corrugation wavelength and the train speed; S2: performing band-pass filtering on the noise and vibration signals inside the car based on the cut-off frequency; S3: identifying the peaks and troughs of the vibration and noise signals after band-pass filtering, and calculating the wave depth index, that is, the vibration and noise characteristics; S4: taking the distance between the peaks as the wavelength of the rail corrugation. In practical applications, the vibration and noise response in the car can be collected through lightweight sensing technology, the data acquisition cost is low, and the data collection method is more convenient and quick, which can meet the needs of rapid measurement of rail corrugation with dense time and space on subway lines.

[0029] In some other embodiments, in step S1, there is the following relationship between the main vibration frequency of the in-vehicle vibration and noise signal and the rail corrugation wavelength: ; In the formula, is the train speed (m / s); is the wavelength of rail corrugation (m), which can be obtained based on existing statistical data; is the main frequency of the noise and vibration signal inside the car (Hz).

[0030] The train speed and rail corrugation wavelength vary, and with other noise interference inside the train, the noise and vibration signals inside the train contain upper and lower cutoff frequencies, namely: ; In the formula, and are the upper and lower cut-off frequencies of the bandpass filter respectively.

[0031] In step S2, the timing noise signal (denoted as ) to perform filtering processing to obtain the characteristics of the noise signal in the corresponding frequency band: ; In the formula, is the noise signal characteristic in a specific frequency band, is the timing noise signal The frequency domain expression of is calculated by Fourier transform; ; It is a bandpass filter, and its characteristics are expressed as follows: ; In the formula, and are the upper and lower cutoff frequencies of the bandpass filter respectively; in actual situations, the upper and lower cutoff frequencies of the bandpass filter can be further determined based on the time-frequency analysis results of the noise signal; the filtered time series noise signal is recorded as , can be obtained by inverse Fourier transform: .

[0032] In step S3, after the noise data is calibrated by mileage, the filtered time series noise signal can be Converted into spatial sequence data along the route ; For the filtered noise spatial sequence To identify the peak value, the peak value must meet the following conditions at the same time: ; In the formula, is the mileage position of the peak value, and They are exist The first and second derivatives at Defined as wave depth index.

[0033] In step S4, after obtaining the peak position, the characteristic wavelength can be estimated based on the mileage difference between adjacent peak positions. ,in, and They are Mileage positions of adjacent crests.

[0034] The following is attached to this application specification Figures 2 to 4 The specific application effects are described as follows: It can be seen that in places where rail corrugation is serious, the wave depth index, a feature extracted from the in-vehicle noise signal, is also relatively large; at the same time, the wavelength estimated based on the wave depth index is basically consistent with the wavelength of rail corrugation measured on site, mainly around 40mm; Rail corrugation identification results of in-vehicle noise wave depth index: (a) Rail corrugation detection results; (b) In-vehicle noise wave depth index; (c) Actual wavelength of rail corrugation; (d) Estimated wavelength; Under the interference of the external environment (such as noise measurement errors, wheel-rail noise, and passenger noise), as the vibration noise wave depth index increases, the wave depth of rail corrugation shows a "fan-shaped" growth, and the in-vehicle noise management standards for rail corrugation maintenance under different reliability levels can be further obtained.

[0035] The technical effects of solving the technical problem in this application are as follows: Compared with the prior art, the present invention provides a train noise and vibration feature extraction method 1 for rail corrugation identification by simultaneously adopting the following steps: S1: determining the cutoff frequency of the noise and vibration inside the car according to the main distribution characteristics of the line corrugation wavelength and the train running speed; S2: performing band-pass filtering on the noise and vibration signals inside the car based on the cut-off frequency; S3: identifying the peaks and troughs of the noise and vibration signals after band-pass filtering, and calculating the wave depth index, i.e., the vibration and noise characteristics; S4: taking the distance between the peaks as the rail corrugation wavelength. In practical applications, the noise and vibration response inside the car can be collected by lightweight sensing technology, the data acquisition cost is low, and the data collection method is more convenient and quick, which can meet the needs of rapid measurement of rail corrugation with dense time and space on subway lines.

[0036] The above-described embodiments of the present invention do not constitute a limitation on the protection scope of the present invention. Any modification, equivalent substitution and improvement made within the spirit and principle of the present invention shall be included in the protection scope of the claims of the present invention.

Claims

1. A train noise feature extraction method for rail corrugation identification, characterized by: The following steps are included: S1: Determine the cutoff frequency of noise and vibration inside the train based on the main distribution characteristics of the line corrugation wavelength and the train running speed; S2: Bandpass filtering of the noise and vibration signals inside the vehicle based on the cutoff frequency; S3: Identify the peaks and valleys of the vibration and noise signal after bandpass filtering, and calculate the wave depth index, i.e., the vibration and noise characteristics; S4: The distance between wave crests is taken as the rail corrugation wavelength.

2. The train noise feature extraction method for rail corrugation identification according to claim 1, characterized in that: In step S1, there is a relationship between the main vibration frequency of the in-vehicle vibration and noise signal and the rail corrugation wavelength as follows: ; In the formula, is the train speed (m / s); is the wavelength of rail corrugation (m), which can be obtained based on existing statistical data; is the main frequency of the noise and vibration signal inside the car (Hz).

3. The train noise feature extraction method for rail corrugation identification according to claim 2, characterized in that: The train speed and rail corrugation wavelength vary, and with other noise interference inside the train, the noise and vibration signals inside the train contain upper and lower cutoff frequencies, namely: ； In the formula, and are the upper and lower cut-off frequencies of the bandpass filter respectively.

4. The train noise feature extraction method for rail corrugation identification according to claim 3, characterized in that: In step S2, the timing noise signal (denoted as ) to perform filtering processing to obtain the characteristics of the noise signal in the corresponding frequency band: ; In the formula, is the noise signal characteristic in a specific frequency band, is the timing noise signal The frequency domain expression of is calculated by Fourier transform; ; It is a bandpass filter, and its characteristics are expressed as follows: ; In the formula, and are the upper and lower cutoff frequencies of the bandpass filter respectively; in actual situations, the upper and lower cutoff frequencies of the bandpass filter can be further determined based on the time-frequency analysis results of the noise signal; the filtered time series noise signal is recorded as , can be obtained by inverse Fourier transform: .

5. The train noise feature extraction method for rail corrugation identification according to claim 1, characterized in that: In step S3, after the noise data is calibrated by mileage, the filtered time series noise signal can be Converted into spatial sequence data along the route ; The filtered noise spatial sequence To identify the peak value, the peak value must meet the following conditions at the same time: ; In the formula, is the mileage position of the peak value, and They are exist The first and second derivatives at Defined as wave depth index.

6. The train noise feature extraction method for rail corrugation identification according to claim 1, characterized in that: In step S4, after obtaining the peak position, the characteristic wavelength can be estimated based on the mileage difference between adjacent peak positions. ,in, and They are Mileage positions of adjacent crests.

Images