CN117932499A - Method for monitoring abnormity of toothed rail - Google Patents
Method for monitoring abnormity of toothed rail Download PDFInfo
- Publication number
- CN117932499A CN117932499A CN202410324695.2A CN202410324695A CN117932499A CN 117932499 A CN117932499 A CN 117932499A CN 202410324695 A CN202410324695 A CN 202410324695A CN 117932499 A CN117932499 A CN 117932499A
- Authority
- CN
- China
- Prior art keywords
- signal
- sub
- abnormal
- vibration
- sound
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01H—MEASUREMENT OF MECHANICAL VIBRATIONS OR ULTRASONIC, SONIC OR INFRASONIC WAVES
- G01H17/00—Measuring mechanical vibrations or ultrasonic, sonic or infrasonic waves, not provided for in the preceding groups
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B61—RAILWAYS
- B61K—AUXILIARY EQUIPMENT SPECIALLY ADAPTED FOR RAILWAYS, NOT OTHERWISE PROVIDED FOR
- B61K9/00—Railway vehicle profile gauges; Detecting or indicating overheating of components; Apparatus on locomotives or cars to indicate bad track sections; General design of track recording vehicles
- B61K9/08—Measuring installations for surveying permanent way
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01H—MEASUREMENT OF MECHANICAL VIBRATIONS OR ULTRASONIC, SONIC OR INFRASONIC WAVES
- G01H1/00—Measuring characteristics of vibrations in solids by using direct conduction to the detector
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/213—Feature extraction, e.g. by transforming the feature space; Summarisation; Mappings, e.g. subspace methods
- G06F18/2131—Feature extraction, e.g. by transforming the feature space; Summarisation; Mappings, e.g. subspace methods based on a transform domain processing, e.g. wavelet transform
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/243—Classification techniques relating to the number of classes
- G06F18/2431—Multiple classes
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/243—Classification techniques relating to the number of classes
- G06F18/2433—Single-class perspective, e.g. one-against-all classification; Novelty detection; Outlier detection
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/25—Fusion techniques
- G06F18/251—Fusion techniques of input or preprocessed data
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/25—Fusion techniques
- G06F18/254—Fusion techniques of classification results, e.g. of results related to same input data
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
- G10L25/00—Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00
- G10L25/03—Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00 characterised by the type of extracted parameters
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
- G10L25/00—Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00
- G10L25/03—Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00 characterised by the type of extracted parameters
- G10L25/18—Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00 characterised by the type of extracted parameters the extracted parameters being spectral information of each sub-band
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
- G10L25/00—Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00
- G10L25/48—Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00 specially adapted for particular use
- G10L25/51—Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00 specially adapted for particular use for comparison or discrimination
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2123/00—Data types
- G06F2123/02—Data types in the time domain, e.g. time-series data
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
- G06F2218/08—Feature extraction
- G06F2218/10—Feature extraction by analysing the shape of a waveform, e.g. extracting parameters relating to peaks
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
- G06F2218/12—Classification; Matching
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- General Physics & Mathematics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- General Engineering & Computer Science (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Evolutionary Computation (AREA)
- Bioinformatics & Computational Biology (AREA)
- Evolutionary Biology (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- Health & Medical Sciences (AREA)
- Acoustics & Sound (AREA)
- Audiology, Speech & Language Pathology (AREA)
- Multimedia (AREA)
- Human Computer Interaction (AREA)
- Signal Processing (AREA)
- Computational Linguistics (AREA)
- Spectroscopy & Molecular Physics (AREA)
- Mechanical Engineering (AREA)
- Measurement Of Mechanical Vibrations Or Ultrasonic Waves (AREA)
Abstract
The invention discloses a rack abnormity monitoring method, which belongs to the technical field of rack abnormity monitoring, wherein a vibration sensor is adopted to collect vibration signals, a broadband sound sensor is adopted to collect sound signals, a vibration signal set and a sound signal set after the whole running process are constructed, signal primary screening is carried out, abnormal vibration signals and abnormal sound signals are screened out, and in order to improve the abnormity monitoring precision, the invention takes a union set for abnormal positions, so that the rack abnormity type classification is carried out according to the vibration signals and the sound signals at each abnormal rack position. Compared with a manual mode, the method has the advantage of real-time monitoring, and based on the sensing data, the method performs fine analysis, so that the accuracy of anomaly monitoring is improved.
Description
Technical Field
The invention relates to the technical field of rack abnormity monitoring, in particular to a rack abnormity monitoring method.
Background
The rack is a specially designed track system for helping trains climb steep slopes. A toothed rail is a special mountain climbing rail that works by adding a toothed track between ordinary rails. Such railways enable trains to climb steeper slopes than normal railways. Locomotives running on the track rail are equipped with one or more gears which mesh with the track rail intermediate the rails. The design solves the problem of insufficient adhesive force of the common wheel track, so that the train can stably run on a steep slope.
The existing method mainly carries out abnormal monitoring on the toothed rail in a manual inspection mode, but the manual inspection effort and precision are limited, the toothed rail is not observed carefully enough, and the problem of low abnormal monitoring precision exists.
Disclosure of Invention
Aiming at the defects in the prior art, the rack abnormality monitoring method provided by the invention solves the problem of low abnormality monitoring precision existing in the existing rack abnormality monitoring method.
In order to achieve the aim of the invention, the invention adopts the following technical scheme: a rack abnormity monitoring method comprises the following steps:
S1, arranging a vibration sensor and a broadband sound sensor at intervals in a toothed rail railway;
s2, splicing vibration signals acquired by the vibration sensor when the train moves to obtain a whole-course vibration signal set;
S3, splicing sound signals acquired by the broadband sound sensor when the train moves to obtain a whole-course sound signal set;
S4, screening abnormal vibration signals for the whole-process vibration signal set, and screening abnormal sound signals for the whole-process sound signal set;
s5, taking a union set of the position of the rack where the abnormal vibration signal is located and the position of the rack where the abnormal sound signal is located, and obtaining an abnormal rack position set;
S6, constructing a classification model according to the vibration signal and the sound signal of each abnormal rack position in the abnormal rack position set, and classifying the abnormal types of the racks.
The beneficial effects of the invention are as follows: in the invention, a vibration sensor is adopted to collect vibration signals, a broadband sound sensor is adopted to collect sound signals, a vibration signal set and a sound signal set after the whole running process are constructed, signal primary screening is carried out, abnormal vibration signals and abnormal sound signals are screened out from the vibration signal set and the sound signal set, and in order to improve the accuracy of abnormal monitoring, the invention takes the union of abnormal positions, thereby classifying the abnormal types of the toothed rails according to the vibration signals and the sound signals at the positions of each abnormal toothed rail. Compared with a manual mode, the method has the advantage of real-time monitoring, and based on the sensing data, the method performs fine analysis, so that the accuracy of anomaly monitoring is improved.
Further, the whole-course vibration signal set or the whole-course sound signal set in S4 is designated as a whole-course signal set, the abnormal vibration signal or the abnormal sound signal is designated as an abnormal signal, and the vibration signal or the sound signal is designated as a sub-signal, and then both screening processes in S4 include the following steps:
S41, calculating a data level factor of the whole-course signal set;
s42, calculating a data level factor of the sub-signals in the whole-process signal set;
s43, calculating the distance between the data level factor of the sub-signal and the data level factor of the whole-process signal set;
and S44, when the distance is greater than the threshold value, the corresponding sub-signal is an abnormal signal.
Further, the formula for calculating the data level factor of the global signal set in S41 is as follows:
，
wherein, O w is the data level factor of the whole course signal set, r i is the ith data value in the whole course signal set, i is a positive integer, and N is the number of data values in the whole course signal set.
Further, the formula for calculating the data level factor of the sub-signals in the global signal set in S42 is as follows:
，
Wherein, O s is the data level factor of the sub-signal, r j is the j-th data value in the sub-signal, j is a positive integer, and M is the number of data values in the sub-signal.
The beneficial effects of the above further scheme are: the invention compares the whole-course data level factor with the data level factor of the sub-signal, and finds out the sub-signal which deviates from the average level obviously as the abnormal signal when the difference is larger than the threshold value, so as to perform the primary screening of the signal.
Further, the calculation formula of the distance in S43 is:
，
Where d is the distance.
Further, the step S6 includes the following sub-steps:
S61, respectively performing time-frequency conversion on vibration signals and sound signals at each abnormal rack position in the abnormal rack position set to obtain vibration frequency domain signals and sound frequency domain signals;
s62, respectively extracting an amplitude mean value, a frequency spectrum dispersion degree and a frequency band position fluctuation degree from the vibration frequency domain signal and the sound frequency domain signal;
S63, inputting the amplitude mean value, the frequency spectrum dispersion and the frequency band position fluctuation degree into a classification model to obtain the rack abnormal type.
The beneficial effects of the above further scheme are: when the rack is aged or has cracks, the vibration frequency and amplitude of the rack and the frequency and amplitude of sound are obviously changed, so that by utilizing the point, the invention performs time-frequency conversion on the vibration signal and the sound signal, and the abnormal type of the rack is obtained by extracting the frequency domain signal characteristics of the two frequency domain signals and processing the frequency domain signal characteristics through a classifier.
Further, in S62, the formulas for extracting the amplitude mean values of the vibration frequency domain signal and the sound frequency domain signal respectively are:
，
wherein, Is the average value of the amplitude/>The amplitude of the nth spectral line in the frequency domain signal is represented by K, the number of the spectral lines is represented by n, and n is a positive integer;
in the step S62, formulas for extracting the frequency spectrum dispersion of the vibration frequency domain signal and the sound frequency domain signal respectively are:
，
wherein, For the frequency spectrum dispersion, f n is the frequency value of the nth spectral line in the frequency domain signal;
the formulas for extracting the frequency band position fluctuation degree from the vibration frequency domain signal and the sound frequency domain signal in the step S62 are as follows:
，
wherein, Is the frequency band position fluctuation degree.
The beneficial effects of the above further scheme are: the invention selects the amplitude mean value to represent the intensity of vibration and sound, selects the spectrum dispersion to represent the main frequency components of vibration and sound, obtains the distribution of the main frequency components, selects the frequency band position fluctuation degree to represent the fluctuation condition of the spectrum, and analyzes the component characteristics of vibration signals and sound signals through the characteristics of three aspects.
Further, the classification model includes: the first input layer, the second input layer, the first sub-classification layer, the second sub-classification layer and the classification output layer;
The input end of the first input layer is used for inputting the amplitude mean value, the frequency spectrum dispersion and the frequency band position fluctuation of the vibration frequency domain signal; the input end of the second input layer is used for inputting the amplitude mean value, the frequency spectrum dispersion and the frequency band position fluctuation of the sound frequency domain signal; the input end of the first sub-classification layer is connected with the output end of the first input layer; the input end of the second sub-classification layer is connected with the output end of the second input layer; the input end of the classification output layer is respectively connected with the output end of the first sub-classification layer and the output end of the second sub-classification layer, and the output end of the classification output layer is used as the output end of the classification model.
Further, the expressions of the first input layer and the second input layer are:
，
Wherein x f,m is the m-th output of the first input layer or the second input layer, the amplitude mean value, the frequency spectrum dispersion and the frequency band position fluctuation degree are constructed into a frequency domain sequence, x c is the mean value of the frequency domain sequence, x v is the variance of the frequency domain sequence, a is a denominator coefficient, x m is the m-th element in the frequency domain sequence, omega in,m is the m-th input weight, b in,m is the m-th input offset, and the m is 1,2 and 3;
The expressions of the first sub-classification layer and the second sub-classification layer are:
，
Where y is the output of the first sub-classification layer or the second sub-classification layer, ω f,m is the mth weight of the sub-classification layer, b f,m is the mth bias of the sub-classification layer, and σ is the sigmoid function.
Further, the expression of the classification output layer is:
，
Where h is the output of the classification model, ln is a logarithmic function, e is a natural constant, y 1 is the output of the first sub-classification layer, and y 2 is the output of the second sub-classification layer.
The beneficial effects of the above further scheme are: according to the invention, two frequency domain sequences are respectively processed through the two input layers, different weights and offsets are given to each element, and then the two sub-classification layers are used for respectively classifying, so that the classification results of the two sub-classification layers are integrated, the integral classification output is obtained, the signal characteristics of the two sensors are integrated, and the classification precision is improved.
Drawings
FIG. 1 is a flow chart of a method for monitoring rack anomalies;
Fig. 2 is a schematic structural diagram of a classification model.
Detailed Description
The following description of the embodiments of the present invention is provided to facilitate understanding of the present invention by those skilled in the art, but it should be understood that the present invention is not limited to the scope of the embodiments, and all the inventions which make use of the inventive concept are protected by the spirit and scope of the present invention as defined and defined in the appended claims to those skilled in the art.
As shown in fig. 1, a rack anomaly monitoring method includes the following steps:
S1, arranging a vibration sensor and a broadband sound sensor at intervals in a toothed rail railway;
s2, splicing vibration signals acquired by the vibration sensor when the train moves to obtain a whole-course vibration signal set;
in the invention, the whole-course vibration signal set is formed by splicing vibration signals acquired by a whole-course vibration sensor;
S3, splicing sound signals acquired by the broadband sound sensor when the train moves to obtain a whole-course sound signal set;
in the invention, the whole-course sound signal set is formed by splicing sound signals acquired by a whole-course broadband sound sensor;
S4, screening abnormal vibration signals for the whole-process vibration signal set, and screening abnormal sound signals for the whole-process sound signal set;
s5, taking a union set of the position of the rack where the abnormal vibration signal is located and the position of the rack where the abnormal sound signal is located, and obtaining an abnormal rack position set;
S6, constructing a classification model according to the vibration signal and the sound signal of each abnormal rack position in the abnormal rack position set, and classifying the abnormal types of the racks.
According to the invention, the plurality of vibration sensors and the plurality of broadband sound sensors are distributed in the rack rail, and a wide frequency range can be covered by the broadband sound sensors, so that sound detection of the rack rail is carried out by adopting the broadband sound sensors, the sound frequency range from tens of kHz to several MHz is better monitored, and the accuracy of monitoring the abnormal rack rail is improved.
The whole-course vibration signal set or the whole-course sound signal set in the step S4 is named as a whole-course signal set, the abnormal vibration signal or the abnormal sound signal is named as an abnormal signal, and the vibration signal or the sound signal is named as a sub-signal, and then the two screening processes in the step S4 comprise the following steps:
S41, calculating a data level factor of the whole-course signal set;
s42, calculating a data level factor of the sub-signals in the whole-process signal set;
s43, calculating the distance between the data level factor of the sub-signal and the data level factor of the whole-process signal set;
and S44, when the distance is greater than the threshold value, the corresponding sub-signal is an abnormal signal.
The formula for calculating the data level factor of the whole course signal set in S41 is as follows:
，
wherein, O w is the data level factor of the whole course signal set, r i is the ith data value in the whole course signal set, i is a positive integer, and N is the number of data values in the whole course signal set.
In this embodiment, when the whole-course signal set is a whole-course vibration signal set, the sub-signals are vibration signals, and the data values are vibration data; when the whole-course signal set is a whole-course sound signal set, the sub-signals are sound signals, and the data values are sound data.
The formula for calculating the data level factor of the sub-signals in the whole signal set in S42 is as follows:
，
Wherein, O s is the data level factor of the sub-signal, r j is the j-th data value in the sub-signal, j is a positive integer, and M is the number of data values in the sub-signal.
The invention compares the whole-course data level factor with the data level factor of the sub-signal, and finds out the sub-signal which deviates from the average level obviously as the abnormal signal when the difference is larger than the threshold value, so as to perform the primary screening of the signal.
The calculation formula of the distance in S43 is:
，
Where d is the distance.
The step S6 comprises the following substeps:
S61, respectively performing time-frequency conversion on vibration signals and sound signals at each abnormal rack position in the abnormal rack position set to obtain vibration frequency domain signals and sound frequency domain signals;
s62, respectively extracting an amplitude mean value, a frequency spectrum dispersion degree and a frequency band position fluctuation degree from the vibration frequency domain signal and the sound frequency domain signal;
S63, inputting the amplitude mean value, the frequency spectrum dispersion and the frequency band position fluctuation degree into a classification model to obtain the rack abnormal type.
When the rack is aged or has cracks, the vibration frequency and amplitude of the rack and the frequency and amplitude of sound are obviously changed, so that by utilizing the point, the invention performs time-frequency conversion on the vibration signal and the sound signal, and the abnormal type of the rack is obtained by extracting the frequency domain signal characteristics of the two frequency domain signals and processing the frequency domain signal characteristics through a classifier.
In the step S62, formulas for extracting the amplitude mean values of the vibration frequency domain signal and the sound frequency domain signal respectively are:
，
wherein, Is the average value of the amplitude/>The amplitude of the nth spectral line in the frequency domain signal is represented by K, the number of the spectral lines is represented by n, and n is a positive integer;
in the step S62, formulas for extracting the frequency spectrum dispersion of the vibration frequency domain signal and the sound frequency domain signal respectively are:
，
wherein, For the frequency spectrum dispersion, f n is the frequency value of the nth spectral line in the frequency domain signal;
the formulas for extracting the frequency band position fluctuation degree from the vibration frequency domain signal and the sound frequency domain signal in the step S62 are as follows:
，
wherein, Is the frequency band position fluctuation degree.
The invention selects the amplitude mean value to represent the intensity of vibration and sound, selects the spectrum dispersion to represent the main frequency components of vibration and sound, obtains the distribution of the main frequency components, selects the frequency band position fluctuation degree to represent the fluctuation condition of the spectrum, and analyzes the component characteristics of vibration signals and sound signals through the characteristics of three aspects.
As shown in fig. 2, the classification model includes: the first input layer, the second input layer, the first sub-classification layer, the second sub-classification layer and the classification output layer;
The input end of the first input layer is used for inputting the amplitude mean value, the frequency spectrum dispersion and the frequency band position fluctuation of the vibration frequency domain signal; the input end of the second input layer is used for inputting the amplitude mean value, the frequency spectrum dispersion and the frequency band position fluctuation of the sound frequency domain signal; the input end of the first sub-classification layer is connected with the output end of the first input layer; the input end of the second sub-classification layer is connected with the output end of the second input layer; the input end of the classification output layer is respectively connected with the output end of the first sub-classification layer and the output end of the second sub-classification layer, and the output end of the classification output layer is used as the output end of the classification model.
The expressions of the first input layer and the second input layer are:
，
Wherein x f,m is the m-th output of the first input layer or the second input layer, the amplitude mean value, the frequency spectrum dispersion and the frequency band position fluctuation degree are constructed into a frequency domain sequence, x c is the mean value of the frequency domain sequence, x v is the variance of the frequency domain sequence, a is a denominator coefficient, x m is the m-th element in the frequency domain sequence, omega in,m is the m-th input weight, b in,m is the m-th input offset, and the m is 1,2 and 3;
The expressions of the first sub-classification layer and the second sub-classification layer are:
，
Where y is the output of the first sub-classification layer or the second sub-classification layer, ω f,m is the mth weight of the sub-classification layer, b f,m is the mth bias of the sub-classification layer, and σ is the sigmoid function.
The expression of the classified output layer is as follows:
，
Where h is the output of the classification model, ln is a logarithmic function, e is a natural constant, y 1 is the output of the first sub-classification layer, and y 2 is the output of the second sub-classification layer.
According to the invention, two frequency domain sequences are respectively processed through the two input layers, different weights and offsets are given to each element, and then the two sub-classification layers are used for respectively classifying, so that the classification results of the two sub-classification layers are integrated, the integral classification output is obtained, the signal characteristics of the two sensors are integrated, and the classification precision is improved.
The larger the output h of the classification model is, the higher the degree of rack abnormality is, and the specific types of rack abnormality can be divided into: mild injury, moderate injury, and severe injury.
In the invention, a vibration sensor is adopted to collect vibration signals, a broadband sound sensor is adopted to collect sound signals, a vibration signal set and a sound signal set after the whole running process are constructed, signal primary screening is carried out, abnormal vibration signals and abnormal sound signals are screened out from the vibration signal set and the sound signal set, and in order to improve the accuracy of abnormal monitoring, the invention takes the union of abnormal positions, thereby classifying the abnormal types of the toothed rails according to the vibration signals and the sound signals at the positions of each abnormal toothed rail. Compared with a manual mode, the method has the advantage of real-time monitoring, and based on the sensing data, the method performs fine analysis, so that the accuracy of anomaly monitoring is improved.
The above is only a preferred embodiment of the present invention, and is not intended to limit the present invention, but various modifications and variations can be made to the present invention by those skilled in the art. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention should be included in the protection scope of the present invention.
Claims (10)
1. The rack abnormity monitoring method is characterized by comprising the following steps:
S1, arranging a vibration sensor and a broadband sound sensor at intervals in a toothed rail railway;
s2, splicing vibration signals acquired by the vibration sensor when the train moves to obtain a whole-course vibration signal set;
S3, splicing sound signals acquired by the broadband sound sensor when the train moves to obtain a whole-course sound signal set;
S4, screening abnormal vibration signals for the whole-process vibration signal set, and screening abnormal sound signals for the whole-process sound signal set;
s5, taking a union set of the position of the rack where the abnormal vibration signal is located and the position of the rack where the abnormal sound signal is located, and obtaining an abnormal rack position set;
S6, constructing a classification model according to the vibration signal and the sound signal of each abnormal rack position in the abnormal rack position set, and classifying the abnormal types of the racks.
2. The method for monitoring the abnormal track according to claim 1, wherein the whole-course vibration signal set or whole-course sound signal set in S4 is designated as a whole-course signal set, the abnormal vibration signal or abnormal sound signal is designated as an abnormal signal, and the vibration signal or sound signal is designated as a sub-signal, and then both screening processes in S4 comprise the steps of:
S41, calculating a data level factor of the whole-course signal set;
s42, calculating a data level factor of the sub-signals in the whole-process signal set;
s43, calculating the distance between the data level factor of the sub-signal and the data level factor of the whole-process signal set;
and S44, when the distance is greater than the threshold value, the corresponding sub-signal is an abnormal signal.
3. The rack anomaly monitoring method according to claim 2, wherein the formula for calculating the data level factor of the global signal set in S41 is:
，
wherein, O w is the data level factor of the whole course signal set, r i is the ith data value in the whole course signal set, i is a positive integer, and N is the number of data values in the whole course signal set.
4. A rack anomaly monitoring method according to claim 3, wherein the formula for calculating the data level factor of the sub-signals in the global signal set in S42 is:
，
Wherein, O s is the data level factor of the sub-signal, r j is the j-th data value in the sub-signal, j is a positive integer, and M is the number of data values in the sub-signal.
5. The rack anomaly monitoring method according to claim 4, wherein the calculation formula of the distance in S43 is:
，
Where d is the distance.
6. The rack anomaly monitoring method according to claim 1, wherein S6 comprises the sub-steps of:
S61, respectively performing time-frequency conversion on vibration signals and sound signals at each abnormal rack position in the abnormal rack position set to obtain vibration frequency domain signals and sound frequency domain signals;
s62, respectively extracting an amplitude mean value, a frequency spectrum dispersion degree and a frequency band position fluctuation degree from the vibration frequency domain signal and the sound frequency domain signal;
S63, inputting the amplitude mean value, the frequency spectrum dispersion and the frequency band position fluctuation degree into a classification model to obtain the rack abnormal type.
7. The method for monitoring abnormal racks as set forth in claim 6, wherein the formulas for extracting the amplitude mean values of the vibration frequency domain signal and the sound frequency domain signal in S62 are:
，
wherein, Is the average value of the amplitude/>The amplitude of the nth spectral line in the frequency domain signal is represented by K, the number of the spectral lines is represented by n, and n is a positive integer;
in the step S62, formulas for extracting the frequency spectrum dispersion of the vibration frequency domain signal and the sound frequency domain signal respectively are:
，
wherein, For the frequency spectrum dispersion, f n is the frequency value of the nth spectral line in the frequency domain signal;
the formulas for extracting the frequency band position fluctuation degree from the vibration frequency domain signal and the sound frequency domain signal in the step S62 are as follows:
，
wherein, Is the frequency band position fluctuation degree.
8. The rack anomaly monitoring method of claim 6, wherein the classification model comprises: the first input layer, the second input layer, the first sub-classification layer, the second sub-classification layer and the classification output layer;
The input end of the first input layer is used for inputting the amplitude mean value, the frequency spectrum dispersion and the frequency band position fluctuation of the vibration frequency domain signal; the input end of the second input layer is used for inputting the amplitude mean value, the frequency spectrum dispersion and the frequency band position fluctuation of the sound frequency domain signal; the input end of the first sub-classification layer is connected with the output end of the first input layer; the input end of the second sub-classification layer is connected with the output end of the second input layer; the input end of the classification output layer is respectively connected with the output end of the first sub-classification layer and the output end of the second sub-classification layer, and the output end of the classification output layer is used as the output end of the classification model.
9. The method for monitoring the abnormal of the rack according to claim 8, wherein the expressions of the first input layer and the second input layer are:
，
Wherein x f,m is the m-th output of the first input layer or the second input layer, the amplitude mean value, the frequency spectrum dispersion and the frequency band position fluctuation degree are constructed into a frequency domain sequence, x c is the mean value of the frequency domain sequence, x v is the variance of the frequency domain sequence, a is a denominator coefficient, x m is the m-th element in the frequency domain sequence, omega in,m is the m-th input weight, b in,m is the m-th input offset, and the m is 1,2 and 3;
The expressions of the first sub-classification layer and the second sub-classification layer are:
，
Where y is the output of the first sub-classification layer or the second sub-classification layer, ω f,m is the mth weight of the sub-classification layer, b f,m is the mth bias of the sub-classification layer, and σ is the sigmoid function.
10. The rack anomaly monitoring method of claim 9, wherein the classification output layer has an expression of:
，
Where h is the output of the classification model, ln is a logarithmic function, e is a natural constant, y 1 is the output of the first sub-classification layer, and y 2 is the output of the second sub-classification layer.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410324695.2A CN117932499B (en) | 2024-03-21 | 2024-03-21 | Method for monitoring abnormity of toothed rail |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410324695.2A CN117932499B (en) | 2024-03-21 | 2024-03-21 | Method for monitoring abnormity of toothed rail |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN117932499A true CN117932499A (en) | 2024-04-26 |
| CN117932499B CN117932499B (en) | 2024-05-31 |
Family
ID=90766692
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202410324695.2A Active CN117932499B (en) | 2024-03-21 | 2024-03-21 | Method for monitoring abnormity of toothed rail |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117932499B (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN118130084A (en) * | 2024-05-08 | 2024-06-04 | 泸州豪能传动技术有限公司 | Differential assembly fault prediction method |
| CN118643298A (en) * | 2024-08-09 | 2024-09-13 | 四川交通职业技术学院 | A method for predicting rack life based on wear identification |
| CN119066401A (en) * | 2024-11-06 | 2024-12-03 | 四川交通职业技术学院 | An early diagnosis method for rack rail cracks using acoustic signal analysis |
Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104236911A (en) * | 2014-09-28 | 2014-12-24 | 江苏润仪仪表有限公司 | Train bogie bearing service process monitoring and fault diagnosis system and method |
| CN110126877A (en) * | 2019-05-21 | 2019-08-16 | 山东交通学院 | A railway track vibration monitoring system and method based on Internet of things technology |
| CN112324629A (en) * | 2020-11-09 | 2021-02-05 | 华能陕西靖边电力有限公司 | Wind power blade early damage monitoring system and method based on vibration and sound |
| CN113763986A (en) * | 2021-09-07 | 2021-12-07 | 山东大学 | A method for detecting abnormal sound of air conditioner indoor unit based on sound classification model |
| CN114462446A (en) * | 2021-12-30 | 2022-05-10 | 上海交通大学 | Rolling bearing fault diagnosis method based on vibration signal and electronic equipment |
| CN115200864A (en) * | 2022-07-05 | 2022-10-18 | 中国国家铁路集团有限公司 | Fault detection method and device for rail vehicle axle box bearing |
| CN116204821A (en) * | 2023-04-27 | 2023-06-02 | 昆明轨道交通四号线土建项目建设管理有限公司 | Vibration evaluation method and system for rail transit vehicle |
| CN116401545A (en) * | 2023-03-29 | 2023-07-07 | 大唐水电科学技术研究院有限公司 | Multimode fusion type turbine runout analysis method |
| CN117009744A (en) * | 2023-07-11 | 2023-11-07 | 安徽徽一通讯科技有限公司 | Ultra-long distance railway sound wave data rapid processing method |
-
2024
- 2024-03-21 CN CN202410324695.2A patent/CN117932499B/en active Active
Patent Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104236911A (en) * | 2014-09-28 | 2014-12-24 | 江苏润仪仪表有限公司 | Train bogie bearing service process monitoring and fault diagnosis system and method |
| CN110126877A (en) * | 2019-05-21 | 2019-08-16 | 山东交通学院 | A railway track vibration monitoring system and method based on Internet of things technology |
| CN112324629A (en) * | 2020-11-09 | 2021-02-05 | 华能陕西靖边电力有限公司 | Wind power blade early damage monitoring system and method based on vibration and sound |
| CN113763986A (en) * | 2021-09-07 | 2021-12-07 | 山东大学 | A method for detecting abnormal sound of air conditioner indoor unit based on sound classification model |
| CN114462446A (en) * | 2021-12-30 | 2022-05-10 | 上海交通大学 | Rolling bearing fault diagnosis method based on vibration signal and electronic equipment |
| CN115200864A (en) * | 2022-07-05 | 2022-10-18 | 中国国家铁路集团有限公司 | Fault detection method and device for rail vehicle axle box bearing |
| CN116401545A (en) * | 2023-03-29 | 2023-07-07 | 大唐水电科学技术研究院有限公司 | Multimode fusion type turbine runout analysis method |
| CN116204821A (en) * | 2023-04-27 | 2023-06-02 | 昆明轨道交通四号线土建项目建设管理有限公司 | Vibration evaluation method and system for rail transit vehicle |
| CN117009744A (en) * | 2023-07-11 | 2023-11-07 | 安徽徽一通讯科技有限公司 | Ultra-long distance railway sound wave data rapid processing method |
Non-Patent Citations (3)
| Title |
|---|
| 王贵: "城轨列车滚动轴承的故障诊断研究", 《中国优秀硕士学位论文全文数据库 (工程科技Ⅱ辑)》, 15 July 2017 (2017-07-15), pages 033 - 333 * |
| 赵宝鹏等: "可搭载智能检测设备的轨道交通单轨自动走行装置研制", 《中国机械》, no. 09, 25 March 2023 (2023-03-25), pages 38 - 41 * |
| 陈静: "基于声音信号分析的牵引变压器故障诊断方法研究", 电气应用, vol. 39, no. 02, 15 February 2020 (2020-02-15), pages 25 - 29 * |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN118130084A (en) * | 2024-05-08 | 2024-06-04 | 泸州豪能传动技术有限公司 | Differential assembly fault prediction method |
| CN118130084B (en) * | 2024-05-08 | 2024-07-05 | 泸州豪能传动技术有限公司 | Differential assembly fault prediction method |
| CN118643298A (en) * | 2024-08-09 | 2024-09-13 | 四川交通职业技术学院 | A method for predicting rack life based on wear identification |
| CN118643298B (en) * | 2024-08-09 | 2024-11-01 | 四川交通职业技术学院 | A method for predicting rack life based on wear identification |
| CN119066401A (en) * | 2024-11-06 | 2024-12-03 | 四川交通职业技术学院 | An early diagnosis method for rack rail cracks using acoustic signal analysis |
Also Published As
| Publication number | Publication date |
|---|---|
| CN117932499B (en) | 2024-05-31 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN117932499B (en) | Method for monitoring abnormity of toothed rail | |
| US6951132B2 (en) | Rail and train monitoring system and method | |
| CN104457643B (en) | A kind of impulse noise filter method and device of track geometry detection data | |
| CN110197588A (en) | A kind of truck driving behavior appraisal procedure and device based on GPS track data | |
| CN113281414B (en) | Method and device for identifying short-wave disease types of steel rails and electronic equipment | |
| CN104833534A (en) | Train running fault diagnosis device based on multi-source information fusion, and method | |
| CN115758289B (en) | A rail corrugation recognition method based on multi-task learning neural network | |
| CN106198062A (en) | The vehicle-mounted acoustic testing system of track vehicle fault and detection method | |
| Dumitriu | Fault detection of damper in railway vehicle suspension based on the cross-correlation analysis of bogie accelerations | |
| CN110879254A (en) | An Acoustic Emission Signal Detection Method for Rail Cracks Based on Improved Least Square Generative Adversarial Network | |
| US20200284671A1 (en) | Method for detecting an anomaly of a rolling equipment exploiting a deformation signal from a rail support | |
| Li et al. | Data-driven structural condition assessment for high-speed railway bridges using multi-band FIR filtering and clustering | |
| CN115204675B (en) | A dynamic risk assessment method and system based on the fusion of multi-source heterogeneous data | |
| CN116297841A (en) | A Method for Identifying Railway Track Diseases Based on Optical Fiber Distributed Vibration Detection | |
| CN104457644B (en) | The detection method and device of non-pulse abnormal data in track geometry detection data | |
| Dai et al. | Acceleration-guided acoustic signal denoising framework based on learnable wavelet transform applied to slab track condition monitoring | |
| CN115371724A (en) | Track state evaluation method and device | |
| CN115905811B (en) | Method for detecting irregularity of railway track height based on FIR-wavelet transformation | |
| CN114777910B (en) | Method for monitoring multi-mode vortex-induced vibration of inhaul cable | |
| CN112113655B (en) | Vibration signal detection device and health assessment method for subway track fastener | |
| CN119066401A (en) | An early diagnosis method for rack rail cracks using acoustic signal analysis | |
| Ghiasi et al. | Monitoring of railway tracks maintenance needs using dynamic responses collected by an in-service train | |
| CN118968733A (en) | A monitoring and early warning system based on array optical fiber sensing technology | |
| CN113032907A (en) | Method and system for correcting vehicle shaking disease data deviation based on waveform correlation | |
| US20200191825A1 (en) | Vehicle detection method |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |