CN113343919B - Method and device for detecting continuous equidistant rubbing damage of steel rail and computer equipment - Google Patents
Method and device for detecting continuous equidistant rubbing damage of steel rail and computer equipment Download PDFInfo
- Publication number
- CN113343919B CN113343919B CN202110744451.6A CN202110744451A CN113343919B CN 113343919 B CN113343919 B CN 113343919B CN 202110744451 A CN202110744451 A CN 202110744451A CN 113343919 B CN113343919 B CN 113343919B
- Authority
- CN
- China
- Prior art keywords
- signal sequence
- scratch
- signal
- value
- unit
- 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.)
- Active
Links
- 229910000831 Steel Inorganic materials 0.000 title claims abstract description 68
- 239000010959 steel Substances 0.000 title claims abstract description 68
- 238000000034 method Methods 0.000 title claims abstract description 33
- 230000006378 damage Effects 0.000 title claims description 31
- 108010076504 Protein Sorting Signals Proteins 0.000 claims abstract description 155
- 238000003466 welding Methods 0.000 claims abstract description 75
- 238000012545 processing Methods 0.000 claims abstract description 41
- 238000001914 filtration Methods 0.000 claims abstract description 32
- 238000001514 detection method Methods 0.000 claims abstract description 26
- 230000001133 acceleration Effects 0.000 claims description 49
- 238000010606 normalization Methods 0.000 claims description 25
- 230000006870 function Effects 0.000 claims description 22
- 238000004590 computer program Methods 0.000 claims description 16
- 238000004364 calculation method Methods 0.000 claims description 9
- 230000001502 supplementing effect Effects 0.000 claims description 9
- 238000013528 artificial neural network Methods 0.000 claims description 8
- 238000013461 design Methods 0.000 claims description 7
- 238000003860 storage Methods 0.000 claims description 7
- 230000004931 aggregating effect Effects 0.000 claims description 6
- 230000009466 transformation Effects 0.000 claims description 5
- 230000008569 process Effects 0.000 claims description 4
- 230000002776 aggregation Effects 0.000 claims description 3
- 238000004220 aggregation Methods 0.000 claims description 3
- 238000000605 extraction Methods 0.000 claims description 3
- 230000011218 segmentation Effects 0.000 claims description 3
- 238000010586 diagram Methods 0.000 description 14
- 238000005516 engineering process Methods 0.000 description 4
- 238000011156 evaluation Methods 0.000 description 3
- 230000003137 locomotive effect Effects 0.000 description 3
- 230000004044 response Effects 0.000 description 3
- 238000012360 testing method Methods 0.000 description 3
- 238000005299 abrasion Methods 0.000 description 2
- 230000007547 defect Effects 0.000 description 2
- 238000002474 experimental method Methods 0.000 description 2
- 230000000737 periodic effect Effects 0.000 description 2
- 238000005070 sampling Methods 0.000 description 2
- 230000001360 synchronised effect Effects 0.000 description 2
- 238000004458 analytical method Methods 0.000 description 1
- 230000008901 benefit Effects 0.000 description 1
- 238000004140 cleaning Methods 0.000 description 1
- 238000004891 communication Methods 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 230000008878 coupling Effects 0.000 description 1
- 238000010168 coupling process Methods 0.000 description 1
- 238000005859 coupling reaction Methods 0.000 description 1
- 238000005336 cracking Methods 0.000 description 1
- 238000007405 data analysis Methods 0.000 description 1
- 238000013500 data storage Methods 0.000 description 1
- 201000010099 disease Diseases 0.000 description 1
- 208000037265 diseases, disorders, signs and symptoms Diseases 0.000 description 1
- 239000006185 dispersion Substances 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 238000007726 management method Methods 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 239000000463 material Substances 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000006855 networking Effects 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 238000006116 polymerization reaction Methods 0.000 description 1
- 238000005201 scrubbing Methods 0.000 description 1
- 238000004901 spalling Methods 0.000 description 1
- 239000000725 suspension Substances 0.000 description 1
- 238000012795 verification Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR 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
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/048—Activation functions
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
- G06F2218/02—Preprocessing
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
- Y02P90/00—Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
- Y02P90/30—Computing systems specially adapted for manufacturing
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Health & Medical Sciences (AREA)
- Software Systems (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Biophysics (AREA)
- Biomedical Technology (AREA)
- Mathematical Physics (AREA)
- Computational Linguistics (AREA)
- Health & Medical Sciences (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Biology (AREA)
- Length Measuring Devices With Unspecified Measuring Means (AREA)
Abstract
The invention discloses a method and a device for detecting continuous equidistant scratch of a steel rail and computer equipment, wherein the method comprises the following steps: acquiring first signal sequences acquired at different positions by a detection vehicle running on a steel rail; identifying and filtering welding joint signals in the first signal sequence, performing resonance demodulation processing on the obtained second signal sequence, and performing standardization processing on the processed third signal sequence to obtain the scratch indexes of different positions of the steel rail; identifying the scratch positions in the steel rail according to preset conditions; the preset condition comprises that the number of continuous peaks exceeding a first threshold value in the scratch index exceeds a second threshold value, and the difference between every two peak intervals is smaller than a preset error distance, wherein each peak interval is the distance between two adjacent continuous peaks. The invention can more accurately detect and identify the continuous equidistant scratch in the steel rail, and solves the problem of lower accuracy of detecting and identifying the continuous equidistant scratch in the related art.
Description
Technical Field
The invention relates to the technical field of rail detection, in particular to a method and a device for detecting continuous equidistant scratch of a steel rail and computer equipment.
Background
The damage of the steel rail is one of the main damage forms of the steel rail and the wheel of the high-speed railway in China at present. The punch damage appears on the tread of the wheel rail, which leads to wheel rail damage and rail irregularity. The sharp points of the punch-damage pits possibly initiate cracks to cause fatigue damage of the wheel rail due to stress concentration. The damage to the steel rail is mainly caused by foreign matters, arc burns, abrasion and block falling. Rail abrasion is a typical disease of high-speed railway rails in China at present. Rail scratches can cause spalling of the rail or lateral fatigue cracking, such as the risk that an untimely treatment will potentially initiate rail breakage. The high-speed railway rail scratches are classified according to sources and can be divided into three types of locomotive scratches before line opening, locomotive group scratches and locomotive scratches after line opening operation. The scratch of the rail may be a defect of the wheel itself or associated with foreign matter attached to the wheel, and thus has a characteristic of continuously appearing equidistantly.
Currently, a method for identifying a damaged position in a rail by an image is provided in the related art. However, the method provided in the related art may not accurately identify scratch due to the problems of the light source, photographing quality, etc. when photographing an image. Moreover, because the length of the steel rail in the shot image is limited, the accuracy of the continuous equidistant scratch identification is low.
Disclosure of Invention
The embodiment of the invention provides a method for detecting continuous equidistant friction injuries of a steel rail, which is used for more accurately detecting and identifying the continuous equidistant friction injuries in the steel rail, so as to solve the technical problem of lower accuracy of detecting and identifying the continuous equidistant friction injuries in the related art. The method comprises the following steps:
acquiring axle box acceleration signals acquired at different positions by a detection vehicle running on a steel rail to obtain a first signal sequence;
identifying and filtering the welding joint signals in the first signal sequence to obtain a second signal sequence;
performing resonance demodulation processing on the second signal sequence to obtain a third signal sequence;
normalizing the third signal sequence to obtain the scratch indexes of different positions of the steel rail;
identifying the scratch positions in the steel rail according to preset conditions; the preset condition comprises that the number of continuous peaks exceeding a first threshold value in the scratch index exceeds a second threshold value, and the difference between every two peak intervals is smaller than a preset error distance, wherein each peak interval is the distance between two adjacent continuous peaks.
Further, performing resonance demodulation processing on the second signal sequence to obtain a third signal sequence, including:
And calculating a Hilbert envelope signal of the second signal sequence through Hilbert transformation to obtain a third signal sequence.
Further, the normalizing process is performed on the third signal sequence to obtain the scratch indexes of different positions of the steel rail, including:
and normalizing the third signal sequence to be within the numerical range of [0,1] by utilizing a sigmoid function of a neural network threshold function to obtain the scratch index.
Further, normalizing the third signal sequence to a value range of [0,1] includes setting the median value of the punch damage index to 1/2 by the following formula:
c＝(x max +x min )/2；
wherein i=1, 2, … N, N is the number of signals of the third signal sequence, W BI,i The scratch index, X of the position corresponding to the ith signal i For the ith signal in the third signal sequence c is the reference value of the signal in the third signal sequence x max Is the maximum value in the third signal sequence, x min For the minimum value in the third signal sequence, a is the first coefficient and a > 0, α is the second coefficient.
Further, identifying the scratch position in the rail according to the preset condition includes:
extracting a local maximum value in the scratch index to obtain a peak value;
calculating the peak distance between every two adjacent peaks;
Dividing the scratch index into a plurality of sections with cross points according to the preset track length; wherein each segment includes a second threshold number of peaks therein;
searching for a section with the same peak value interval in the sections;
and under the condition that all peaks in the searched section exceed a first threshold value, determining that the corresponding section has scratch.
Further, identifying and filtering the welding joint signal in the first signal sequence to obtain a second signal sequence, including:
calculating the moving effective value of each signal in the first signal sequence by the following formula to obtain { RMS } i ,i＝1,2,…N}：
The length of a moving window for moving the effective value is K signals;
segmenting the mobile effective value to obtain a plurality of segments;
according to the average value of each segmentAnd variance sigma, a reference threshold R for each segment is calculated by the following formula T ：
Find exceeding the reference threshold R T Is recorded as a first set of mobile effective valuesWherein N is R Is->Is the number of (3);
aggregating the first set, and keeping the maximum value in the movement effective values with the distance smaller than the preset distance to obtain a second setWherein (1)>Is->Is the number of (3);
determining the position corresponding to each data in the second set to obtain the position of the welding joint;
Supplementing missing welding joint positions in the second set according to the design spacing of the welding joints;
determining a signal corresponding to the welding joint position in the first signal sequence to obtain a welding joint signal;
and setting the K signals near each welding joint signal to a preset value.
The embodiment of the invention also provides a device for detecting the continuous equidistant friction injuries of the steel rail, which is used for more accurately detecting and identifying the continuous equidistant friction injuries in the steel rail so as to solve the technical problem of lower accuracy in detecting and identifying the continuous equidistant friction injuries in the related art, and comprises the following steps:
the acquisition unit is used for acquiring axle box acceleration signals acquired by a detection vehicle running on a steel rail at different positions to obtain a first signal sequence;
the identification filtering unit is used for identifying and filtering the welding joint signals in the first signal sequence to obtain a second signal sequence;
a resonance demodulation unit for performing resonance demodulation processing on the second signal sequence to obtain a third signal sequence;
the normalization processing unit is used for performing normalization processing on the third signal sequence to obtain scratch indexes of different positions of the steel rail;
The identifying unit is used for identifying the scratch position in the steel rail according to preset conditions; the preset condition comprises that the number of continuous peaks exceeding a first threshold value in the scratch index exceeds a second threshold value, and the difference between every two peak intervals is smaller than a preset error distance, wherein each peak interval is the distance between two adjacent continuous peaks.
Further, the resonance demodulation unit is further configured to calculate a hilbert envelope signal of the second signal sequence through hilbert transformation, so as to obtain a third signal sequence.
Further, the normalization processing unit is further configured to normalize the third signal sequence to a numerical range of [0,1] by using a sigmoid function that is a neural network threshold function, so as to obtain a scratch index.
Further, the normalization processing unit is further configured to set the intermediate value of the scratch index to 1/2 by the following formula:
c＝(x max +x min )/2；
wherein i=1, 2, … N, N is the number of signals of the third signal sequence, W BI,i The scratch index, X of the position corresponding to the ith signal i For the ith signal in the third signal sequence c is the reference value of the signal in the third signal sequence x max Is the maximum value in the third signal sequence, x min For the minimum value in the third signal sequence, a is the first coefficient and a > 0, α is the second coefficient.
Further, the identification unit includes:
the extraction unit is used for extracting a local maximum value in the punch damage index to obtain a peak value;
a first calculation unit for calculating a peak interval between every two adjacent peaks;
the dividing unit is used for dividing the scratch index into a plurality of sections with cross according to the preset track length; wherein each segment includes a second threshold number of peaks therein;
the searching unit is used for searching the sections with the same peak value interval in the sections;
and the first determining unit is used for determining that the corresponding section has a scratch under the condition that all peaks in the searched section exceed a first threshold value.
Further, the identifying and filtering unit includes:
a second calculation unit for calculating the movement effective value of each signal in the first signal sequence by the following formula to obtain { RMS } i ,i＝1,2,…N}：
The length of a moving window for moving the effective value is K signals;
the segmentation unit is used for segmenting the mobile effective value to obtain a plurality of segments;
a third calculation unit for calculating an average value of each segmentAnd variance sigma, a reference threshold R for each segment is calculated by the following formula T ：
A searching unit for searching for a reference threshold R T Is recorded as a first set of mobile effective valuesWherein N is R Is->Is the number of (3);
an aggregation unit for aggregating the first set, and keeping the maximum value in the movement effective value smaller than the preset distance to obtain a second setWherein (1)>Is->Is the number of (3);
the second determining unit is used for determining the position corresponding to each data in the second set to obtain the position of the welding joint;
the supplementing unit is used for supplementing the missing welding joint positions in the second set according to the design interval of the welding joints;
the third determining unit is used for determining a signal corresponding to the welding joint position in the first signal sequence to obtain a welding joint signal;
and the setting unit is used for setting the welding joint signal and K signals nearby as preset values.
The embodiment of the invention also provides computer equipment, which comprises a memory, a processor and a computer program stored on the memory and capable of running on the processor, wherein the processor realizes the method for detecting the continuous equidistant rubbing damage of the steel rail when executing the computer program.
The embodiment of the invention also provides a computer readable storage medium, which stores a computer program for executing the method for detecting the continuous equidistant rubbing damage of the steel rail.
In the embodiment of the invention, the first signal sequence is obtained by acquiring axle box acceleration signals acquired at different positions by a detection vehicle running on a steel rail, then a welding joint signal in the first signal sequence is identified and filtered to obtain the second signal sequence, then resonance demodulation processing is carried out on the second signal sequence to obtain the third signal sequence, and normalization processing is carried out on the third signal sequence to obtain the scratch indexes of different positions of the steel rail, so that the scratch positions can be identified in the steel rail according to preset conditions, wherein the preset conditions comprise that the number of continuous peaks exceeding a first threshold value in the scratch indexes exceeds a second threshold value, and the difference between every two peak intervals is smaller than a preset error distance, and each peak interval is the distance between two adjacent continuous peaks. According to the embodiment of the invention, the equidistant scratch indexes obtained through the identification and filtering of the welding joint signals and the resonance demodulation treatment and the normalization treatment can be used for more accurately detecting and identifying the equidistant scratch continuously existing in the steel rail, so that the technical problem of lower accuracy of detecting and identifying the continuous equidistant scratch in the related technology is solved.
Drawings
In order to more clearly illustrate the embodiments of the invention or the technical solutions in the prior art, the drawings that are required in the embodiments or the description of the prior art will be briefly described, it being obvious that the drawings in the following description are only some embodiments of the invention, and that other drawings may be obtained according to these drawings without inventive effort for a person skilled in the art. In the drawings:
FIG. 1 is a schematic flow chart of an alternative method for detecting continuous equidistant wiping of a rail in an embodiment of the invention;
FIG. 2 is a schematic diagram of an alternative configuration of a multi-section vehicle dynamic detection system in accordance with an embodiment of the present invention;
FIG. 3 is a waveform diagram of an exemplary embodiment for calculating a punch damage index;
FIG. 4 is a schematic flow chart of another alternative method for detecting continuous equidistant wiping of a rail in an embodiment of the invention;
FIG. 5 is a waveform diagram showing an example of a signal identifying a weld joint by a method for detecting continuous equidistant wiping of a rail in an embodiment of the invention;
FIG. 6 is a schematic waveform diagram of an axle box acceleration signal acquired at an experimental road section;
FIG. 7 is a waveform diagram of a scratch index obtained by calculating the axle box acceleration shown in FIG. 6 according to a method for detecting continuous equidistant scratch of a rail in an embodiment of the present invention;
FIG. 8 is a block diagram showing an alternative construction of a device for detecting continuous equidistant wiping of a rail in an embodiment of the invention.
Detailed Description
For the purpose of making the objects, technical solutions and advantages of the embodiments of the present invention more apparent, the embodiments of the present invention will be described in further detail with reference to the accompanying drawings. The exemplary embodiments of the present invention and their descriptions herein are for the purpose of explaining the present invention, but are not to be construed as limiting the invention.
The method for detecting the continuous equidistant scratch of the steel rail provided by the embodiment of the invention, as shown in fig. 1, can comprise the following steps:
and 101, acquiring axle box acceleration signals acquired by a detection vehicle running on a steel rail at different positions to obtain a first signal sequence.
In the detection vehicle running on the high-speed railway, the axle box is directly connected with the wheel pair, so that vibration generated by irregularity of the track can be directly transmitted to the axle box through the wheel pair. Thus, the vibratory acceleration of the axle housing may be directly responsive to track shortwave irregularities, and the axle housing acceleration signal may include a vertical vibratory acceleration signal of the axle housing and/or a lateral vibratory acceleration signal of the axle housing. During acquisition, axle box acceleration signals when the vehicle runs at different positions of the steel rail are obtained according to a certain sampling frequency. For example, the axle box acceleration signal can be acquired at a sampling frequency of h Hz on a road section of a high-speed railway to obtain a signal sequence.
The detection of the axle box acceleration signal can be detected by a multi-section vehicle dynamic detection system. An exemplary multi-section vehicle dynamic detection system is shown in fig. 2, and can collect acceleration of a vehicle body, a framework and axle boxes in real time for assisting in analyzing the smoothness state of a turnout. The system adopts a multichannel distributed networking test technology, uses a computer to remotely control test equipment distributed at different places to synchronously work, transmits data and synchronous information through a network, and has the characteristics of large measured data quantity, regional dispersion, high real-time performance and reliability of test, remote cooperative operation and the like. The system has the functions of collecting and processing original signals on line, storing intermediate data and final results, displaying waveform diagrams on line, transmitting data through a network, outputting an overrun report, correcting mileage, playing back the stored data afterwards, outputting waveform diagram data, corresponding places and speeds and the like. And realizing data acquisition, original data storage, data validity judgment and waveform display.
As shown in fig. 2, the multi-section vehicle dynamic detection system (also referred to as an acceleration detection system) includes three section data acquisition devices: the data acquisition device 207, the data acquisition device 208, and the data acquisition device 209 CAN acquire axle box acceleration, frame acceleration, and vehicle body acceleration of the corresponding section, and CAN receive CAN (controller area network ) input data transmitted by the reflective memory card 204, the reflective memory card 205, and the reflective memory card 206, respectively. As shown in fig. 2, the communication network of the acceleration detection system includes a backbone data network 201 and a non-backbone data network 202, signals of three sections can be synchronized by the backbone data network 201, and three reflective memory cards 204, 205, 206 are all connected and communicated with the central control computer 203, and can be used to receive synchronization packets from the integrated system.
Step 102, identifying and filtering the welding joint signal in the first signal sequence to obtain a second signal sequence.
Because the welding joint has certain irregularity, vibration generated by the irregularity is superposed in the axle box acceleration signal, which is equivalent to a frequency modulation and amplitude modulation signal generated when a vehicle passes through the welding joint, in order to eliminate the influence of the welding joint signal on the scratch identification, the welding joint signal in the first signal sequence can be identified and filtered.
Because the welding joints are generally equidistant on the railway, the axle box acceleration signals (namely the welding joint signals) at the welding joints can be automatically identified by using an equidistant energy extremum method, and whether the axle box acceleration signals are the welding joint signals or not is determined according to the equidistant of the local maxima of the signal energy; further, the weld joint signal is filtered out of the first signal sequence to obtain a second signal sequence.
And 103, performing resonance demodulation processing on the second signal sequence to obtain a third signal sequence.
Since axle box acceleration is the result of dynamic coupling of the wheel track, and exhibits high frequency and highly non-linear characteristics. Besides short wave irregularity of the rail such as rail scratch, the shape and material of the wheels and the rail tread, the suspension parameters of the vehicle, the mounting position of the speed sensor and the like have great influence on the acceleration of the axle box. The axle box acceleration amplitude is directly used for diagnosing the rail scratch, so that the problems of high randomness of the judging result and difficult determination of the threshold value can occur.
Therefore, the embodiment of the invention provides a new dynamic evaluation Index for diagnosing the continuous equidistant scratch of the steel rail, namely the scratch Index (WBI). The scratch index is defined as a normalized value of an envelope signal obtained by resonance demodulation of the axle box vibration acceleration.
Compared with the traditional evaluation method taking the dynamic response amplitude of the vehicle as an index, the scratch index can utilize the resonance demodulation envelope signal of the dynamic response of the vehicle to replace the original waveform signal, and the non-stationary high-frequency signal is demodulated into the low-frequency signal with high stability under the condition of not losing the vibration characteristic, so that the problem of high randomness of the detection result is solved; meanwhile, a proper demodulation range is selected according to the modulation characteristics to calculate corresponding indexes, and normalization processing is carried out by combining a large amount of historical detection data, so that the problem that the absolute threshold is difficult to determine is solved.
In calculating the scratch index, it is first necessary to perform resonance demodulation processing on the second signal sequence. In an alternative embodiment, the third signal sequence may be obtained by calculating the hilbert envelope signal of the second signal sequence by hilbert transformation.
To calculate a more accurate punch damage index, in one example, the second signal sequence may be bandpass filtered to generate a fourth signal sequence, and an envelope signal of the fourth signal sequence may be calculated to obtain a third signal sequence. The frequency of the band-pass filtering can be preset in segments, and each segment respectively presets the corresponding band-pass filtering frequency range F L ,F H ]。
And 104, performing normalization processing on the third signal sequence to obtain the scratch indexes of different positions of the steel rail.
Normalization is a means of data processing. Since different evaluation indexes often have different dimensions, the difference in numerical value may be large, and the result of data analysis may be affected without processing. In order to eliminate the influence of the difference between the dimensions and the value ranges of the indexes, the standardization processing is needed, and the data is scaled according to the proportion so as to fall into a specific area, thereby being convenient for comprehensive analysis. Illustratively, the sigmoid function may be normalized to within [0,1] range by the neural network threshold function.
The following provides an alternative embodiment to describe steps 103-104 in detail, which may include the following steps:
the first step, according to the given continuous equidistant scrubbing modulation range of the steel rail, the actually measured axle box acceleration signal (namely the second signal sequence) is subjected to segmented bandpass filtering with the filtering frequency of[F L ,F H ]The axle box acceleration after band-pass filtering is recorded as { x } F,i I=1, 2, … N } (i.e., fourth signal sequence).
Second, calculating axle box acceleration { x after band-pass filtering F,i The Hilbert envelope signal of i=1, 2, … N } (i.e. the fourth signal sequence), denoted as { X } i I=1, 2, … N } (i.e., the third signal sequence).
Third, the envelope signal { X }, is processed i I=1, 2, … N } (i.e. third signal sequence) is normalized to [0,1 ] using the neural network threshold function sigmoid function]Within the range and is described as
Wherein i=1, 2, … N, N is the number of signals of the third signal sequence, W BI,i The scratch index, X of the position corresponding to the ith signal i For the ith signal in the third signal sequence, c is the reference value for the signal in the third signal sequence. W in the above BI For convenience of standardization management, dimension of the characteristic index is unified, measuring range is set to be 1, namely, the number in real space is mapped to [0,1 ]]And (3) inner part.
Thus, the resonance demodulation signal of the axle box acceleration is normalized by the sigmoid function to obtain the scratch index W BI It can be noted that:
W BI ＝sigmoid(x i '), i=1, 2, N formula 2
In the above, x i The' is an envelope signal (a third signal sequence) obtained by resonance demodulation of an axle box acceleration signal (a second signal sequence) at an ith measuring point, and N is the number of signals in each of the first signal sequence to the third signal sequence.
And 105, identifying the scratch position in the steel rail according to preset conditions.
The preset condition comprises that the number of continuous peaks exceeding a first threshold exceeds a second threshold in the scratch index, and the difference between every two peak intervals is smaller than a preset error distance, wherein each peak interval is the distance between two adjacent continuous peaks. That is, when the scratch position is identified, it is necessary to determine a position satisfying the above-described preset condition.
The number of consecutive peaks exceeding the first threshold in the scratch index exceeds a second threshold, i.e. the number of consecutive peaks exceeds a second value, and each consecutive peak exceeds the first threshold, the first threshold being a threshold set for the scratch index, the second threshold being a threshold set for the number of peaks, an exemplary preset condition may be specifically set to include: the peak value of the continuous over 5 punch indices is greater than 0.75.
The difference between every two peak intervals is smaller than the preset error distance, wherein each peak interval is the distance between every two adjacent continuous peaks, that is, the distances between every two adjacent continuous peaks should be consistent, in order to leave a certain error space, the preset error distance can be set, and it is only necessary to determine that the difference between every two peak intervals is smaller than the preset error distance.
Specifically, step 105 of identifying the scratch position in the rail according to the preset condition may include the following steps (1) to (5):
(1) And extracting a local maximum value in the scratch index to obtain a peak value.
(2) The peak spacing between every two adjacent peaks is calculated.
(3) The punch damage index is divided into a plurality of sections with intersections according to the preset track length, wherein each section comprises a second threshold number of peaks.
(4) And searching for the sections with the same peak pitch in the sections.
(5) And under the condition that all peaks in the searched section exceed a first threshold value, determining that the corresponding section has scratch.
Illustratively, when W, with a continuous multimodal number of more than 5 BI When the calculated value is close to 0, the track smoothness is basically good. When W is BI When the calculated value is close to 0.25, the steel rail is slightly and continuously scratched at equal intervals; when W is BI When the calculated value is close to 0.5, the rail is provided with moderate continuous equidistant scratch; when W is BI When the calculated value is close to 0.75, the continuous equidistant scratch degree of the steel rail needs to be checked in time. In particular if W BI Exceeding 0.9, approaching 1. At this time, the continuous equidistant rubbing degree of the steel rail must be checked immediately to ensure the driving safety.
In the embodiment of the invention, the first signal sequence is obtained by acquiring axle box acceleration signals acquired at different positions by a detection vehicle running on a steel rail, then a welding joint signal in the first signal sequence is identified and filtered to obtain the second signal sequence, then resonance demodulation processing is performed on the second signal sequence to obtain the third signal sequence, and normalization processing is performed on the third signal sequence to obtain the scratch indexes of different positions of the steel rail, so that the scratch positions can be identified in the steel rail according to preset conditions, wherein the preset conditions comprise that the number of continuous peaks exceeding a first threshold value in the scratch indexes exceeds a second threshold value. According to the embodiment of the invention, the scratch indexes obtained through the identification and filtering of the welding joint signals and the resonance demodulation treatment and the normalization treatment can be used for more accurately detecting and identifying the scratch in the steel rail, so that the technical problem of lower accuracy in detecting and identifying the scratch in the related technology is solved.
In an alternative embodiment, step 104 normalizes the third signal sequence to within the value range of [0,1], which may include setting the median value of the punch damage index to 1/2 by equation 1 and the following equation:
c＝(x max +x min ) Formula 3/2
Wherein c is a reference value of the signal in the third signal sequence, x max Is the maximum value in the third signal sequence, x min For the minimum value in the third signal sequence, a is the first coefficient and a > 0, α is the second coefficient.
When step 102 is performed to identify and filter the welding joint signal in the first signal sequence to obtain the second signal sequence, if the scratch index and the number of continuous multimodal values are calculated and compared with the threshold value, an exemplary result is shown in fig. 3, and the circled position determines that the steel rail has the scratch characteristic according to the scratch index, but this is a false scratch, which is a false report caused by the influence of the welding joint, so that the welding joint can be automatically identified and filtered by an equal interval energy extremum method, and the scratch false report at the welding joint is eliminated.
Specifically, step 102 identifies and filters the welding joint signal in the first signal sequence to obtain the second signal sequence, and may include the following steps (1) - (9):
(1) Calculating the moving effective value of each signal in the first signal sequence by the following formula to obtain { RMS } i ,i＝1,2,…N}：
Wherein the length of the moving window for moving the effective value is K signals.
(2) And segmenting the mobile effective value to obtain a plurality of segments.
(3) According to the average value of each segmentAnd variance sigma, a reference threshold R for each segment is calculated by the following formula T ：
(4) Find exceeding the reference threshold R T Is recorded as a first set of mobile effective valuesWherein N is R Is->Is a number of (3).
(5) Aggregating the first set, and keeping the maximum value in the movement effective values with the distance smaller than the preset distance to obtain a second setWherein (1)>Is->Is a number of (3).
(6) And determining the position corresponding to each data in the second set to obtain the position of the welding joint.
(7) And supplementing missing welding joint positions in the second set according to the design spacing of the welding joints.
(8) And determining a signal corresponding to the welding joint position in the first signal sequence to obtain a welding joint signal.
(9) And setting the K signals near each welding joint signal to a preset value.
An alternative embodiment of the invention for detecting continuous equidistant cleaning of rails is provided below, and a schematic flow chart of the alternative embodiment is shown in fig. 4. The specific implementation flow can comprise the following steps:
(1) And installing acceleration detection systems of two sections on the high-speed comprehensive detection vehicle, and collecting the actually measured acceleration signals of the axle box to obtain a first signal sequence.
And then, identifying and filtering the welding joint signals in the acceleration signals of the shaft box to obtain a second signal sequence.
(2) After the axle box acceleration data is acquired, the axle box acceleration is bandpass filtered.
The filtering frequency can be taken as [20,500 ]]Hz, the axle box acceleration signal after the filtering is recorded asWhere N is the number of sample points.
(3) And identifying the welded joint. Specifically, the steps 1) to 4) below may be included:
step 1), calculating the axle box acceleration { x } i The effective value of the movement of i=1, 2, … N } is denoted as { RMS } i I=1, 2, … N, wherein,where the length of the moving window for moving the effective value is K signals.
Step 2), segmenting the effective value, and calculating the average value of each segmentAnd variance sigma, calculating a segment reference threshold R of the movement effective value of the axle box acceleration according to the following formula T ，/>
Step 3), finding out that the effective value is larger than or equal to the reference threshold value R T Is defined as a large value overrun welded joint, is noted asWherein N is R The number of the large-value overrun welding joints is the number.
Step 4), pairPolymerizing, wherein the maximum value is reserved for the large value overrun point at the same welding joint, and the polymerized large value overrun welding joint is marked as +. > The number of the large-value overrun welding joints after polymerization; and, according to the design that the intervals of the welding joints are equal, the supplementary effective value is smaller than the reference threshold value R T Is provided. The periodic spacing of the welding joints of the ballastless line is 100m, and the welding joints of the ballastless lineThe periodic spacing is 25m, and welding joints with the spacing of 25m are automatically extracted from a certain ballasted line K18-K18+500.
(4) And filtering out the welding joint signal. That is, the welded joint signal is subjected to a filter process. If i w Is a welded joint point, let x j ＝0,j＝i w -K,i w -K+1,…i w +K-1,i w And +K, wherein K is the number of signals contained in the length of the moving window and is a preset value. The circled peak points in fig. 5 are weld joint locations identified by the equally spaced energy extremum method.
(5) And performing resonance demodulation on the signal after the welding joint signal is filtered.
Firstly, according to the given continuous equidistant friction damage modulation range of the steel rail, the actually measured acceleration of the axle box is subjected to sectional bandpass filtering with the filtering frequency of [ F ] L ,F H ]The axle box acceleration after band-pass filtering is recorded as { x } F,i I=1, 2, … N }, and then calculating the bandpass filtered axlebox acceleration { x } F,i The Hilbert envelope signal of i=1, 2, … N } is denoted as { X } i ,i＝1,2,…N}。
(6) The scratch index WBI is calculated.
The envelope signal { X }, is applied to i I=1, 2, … N } is normalized to [0,1 ] using the neural network threshold function sigmoid function]Within the range and expressed as equation 1.
In order to ensure that the intermediate value of the component can fall at 1/2 (the intermediate value falls to 0.5 after normalization), let c take the value of the average value of the maximum value and the minimum value of the component, namely, formula 3. Accordingly, in order to ensure that the minimum and maximum values of the components can fall at 0 and 1, respectively, and that the increment of the components, a must be greater than 0, and that the value is determined jointly by the maximum and minimum values, equation 4, where α=0.001, maximum x max And a minimum value x min Is a statistical value.
(7) Judging whether the scratch index is larger than a threshold gamma.
(8) It is determined whether or not a plurality of peaks are consecutive greater than a threshold gamma.
(9) And if the seventh step and the eighth step are both yes, determining that the continuous rail scratch exists.
One obvious feature of the continuous equidistant rubbing of the rail is that there are multiple rubbing sites on the surface of the rail in a section. The scratch distance is related to the number of foreign objects attached to the wheels of the train or defects on the wheels, and is generally equal in distance, and the length of the equal distance can be the circumference of the wheels. Taking the case of attaching a foreign matter to the wheel, the rubbing distance was 2.7 m when the diameter of the wheel was 860 mm. According to this feature, the detailed procedure for evaluating continuous multi-peak rail continuous equidistant scratch may include the following steps 1) to 6):
Step 1) finding out the damage index W of the steel rail BI Local maxima of the index;
step 2) calculating the distance between local maxima;
step 3) dividing the input scratch index data and the corresponding mileage into crossed sections according to a specified length;
step 4) judging whether the distances among the local maxima of each section are consistent;
step 5) if the values are consistent, judging that the scratch section exists according to the continuous multimodal quantity threshold value and the scratch index, and outputting the scratch distance, the start and end mileage of the scratch section and the number of points with scratches;
and 6) if the two rails are inconsistent, no rail scratch exists.
Experimental verification was performed by the above detailed description. In one experiment, the measured axle box acceleration of the train was measured at high speed as shown in FIG. 6, and the scratch index was shown in FIG. 7. As can be seen from FIG. 7, the peak value characteristic of the scratch index is obvious, the dynamic response characteristic of the axle box acceleration caused by the scratch of the steel rail can be better depicted, and the impact characteristic is stronger from the original waveform of the signal.
In another experiment, the punch damage index was set to 0.75 and the number of consecutive multi-peaks was set to 5. The calculated scratch index for the uplink K655+300 of a certain high-speed line is 0.99, the scratch distance is 2.7m, the number of continuous multimodal values is 23, the axle box vibration is strong, and the condition that continuous equidistant scratch exists on the steel rail is diagnosed. By rechecking the actual conditions on site, the method for determining the number of the continuous multi-peak values and the combined scratch indexes of the embodiment of the invention can effectively diagnose the continuous equidistant scratch of the steel rail.
The embodiment of the invention also provides a device for detecting the continuous equidistant scratch of the steel rail, as described in the following embodiment. The principle of the device for solving the problem is similar to that of a detection method for continuously and equidistantly rubbing the steel rail, so that the implementation of the device can be referred to the implementation of the detection method for continuously and equidistantly rubbing the steel rail, and the repetition is not repeated.
As shown in fig. 8, the detection device for the continuous equidistant rubbing of the steel rail can comprise an acquisition unit 11, an identification filtering unit 12, a resonance demodulation unit 13, a normalization processing unit 14 and an identification unit 15.
The acquisition unit 11 is configured to acquire axle box acceleration signals acquired at different positions by a detected vehicle traveling on a rail, and obtain a first signal sequence.
The identification and filtering unit 12 is used for identifying and filtering the welding joint signal in the first signal sequence to obtain a second signal sequence.
The resonance demodulating unit 13 is configured to perform resonance demodulation processing on the second signal sequence to obtain a third signal sequence.
The normalization processing unit 14 is configured to perform normalization processing on the third signal sequence to obtain scratch indexes of different positions of the steel rail;
the identifying unit 15 is used for identifying the scratch position in the steel rail according to preset conditions; the preset condition comprises that the number of continuous peaks exceeding a first threshold value in the scratch index exceeds a second threshold value, and the difference between every two peak intervals is smaller than a preset error distance, wherein each peak interval is the distance between two adjacent continuous peaks.
Optionally, the resonance demodulation unit is further configured to calculate a hilbert envelope signal of the second signal sequence by hilbert transformation, to obtain a third signal sequence.
Optionally, the normalization processing unit is further configured to normalize the third signal sequence to a value range of [0,1] by using a sigmoid function that is a neural network threshold function, so as to obtain the scratch index.
Optionally, the normalization processing unit is further configured to set the median value of the scratch index to 1/2 by the following formula:
c＝(x max +x min )/2；
wherein i=1, 2, … N, N is the number of signals of the third signal sequence, W BI,i The scratch index, X of the position corresponding to the ith signal i For the ith signal in the third signal sequence c is the reference value of the signal in the third signal sequence x max Is the maximum value in the third signal sequence, x min For the minimum value in the third signal sequence, a is the first coefficient and a > 0, α is the second coefficient.
Alternatively, the identification unit may include:
the extraction unit is used for extracting a local maximum value in the punch damage index to obtain a peak value;
a first calculation unit for calculating a peak interval between every two adjacent peaks;
the dividing unit is used for dividing the scratch index into a plurality of sections with cross according to the preset track length; wherein each segment includes a second threshold number of peaks therein;
The searching unit is used for searching the sections with the same peak value interval in the sections;
and the first determining unit is used for determining that the corresponding section has a scratch under the condition that all peaks in the searched section exceed a first threshold value.
Optionally, the identifying and filtering unit may include:
a second calculation unit for calculating the difference between the two data byCalculating the movement effective value of each signal in the first signal sequence according to the formula to obtain { RMS } i ,i＝1,2,…N}：
The length of a moving window for moving the effective value is K signals;
the segmentation unit is used for segmenting the mobile effective value to obtain a plurality of segments;
a third calculation unit for calculating an average value of each segmentAnd variance sigma, a reference threshold R for each segment is calculated by the following formula T ：
/>
A searching unit for searching for a reference threshold R T Is recorded as a first set of mobile effective valuesWherein N is R Is->Is the number of (3);
an aggregation unit for aggregating the first set, and keeping the maximum value in the movement effective value smaller than the preset distance to obtain a second setWherein (1)>Is->Is the number of (3);
the second determining unit is used for determining the position corresponding to each data in the second set to obtain the position of the welding joint;
The supplementing unit is used for supplementing the missing welding joint positions in the second set according to the design interval of the welding joints;
the third determining unit is used for determining a signal corresponding to the welding joint position in the first signal sequence to obtain a welding joint signal;
and the setting unit is used for setting the welding joint signal and K signals nearby as preset values.
In the embodiment of the invention, the axle box acceleration signals acquired at different positions by a detection vehicle running on a steel rail are acquired to obtain a first signal sequence, then welding joint signals in the first signal sequence are identified and filtered to obtain a second signal sequence, resonance demodulation processing is further carried out on the second signal sequence to obtain a third signal sequence, normalization processing is carried out on the third signal sequence to obtain the scratch indexes of different positions of the steel rail, and therefore continuous equidistant scratch positions can be identified in the steel rail according to preset conditions, wherein the preset conditions comprise that the number of continuous peaks exceeding a first threshold value in the scratch indexes exceeds a second threshold value, and the difference between every two peak intervals is smaller than a preset error distance, and each peak interval is the distance between two adjacent continuous peaks. According to the embodiment of the invention, the equidistant scratch indexes obtained through the identification and filtering of the welding joint signals and the resonance demodulation treatment and the normalization treatment can be used for more accurately detecting and identifying the equidistant scratch continuously existing in the steel rail, so that the technical problem of lower accuracy of detecting and identifying the continuous equidistant scratch in the related technology is solved.
The embodiment of the invention also provides computer equipment, which comprises a memory, a processor and a computer program stored on the memory and capable of running on the processor, wherein the processor realizes the method for detecting the continuous equidistant rubbing damage of the steel rail when executing the computer program.
The embodiment of the invention also provides a computer readable storage medium, which stores a computer program for executing the method for detecting the continuous equidistant rubbing damage of the steel rail.
It will be appreciated by those skilled in the art that embodiments of the present invention may be provided as a method, system, or computer program product. Accordingly, the present invention may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present invention may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, and the like) having computer-usable program code embodied therein.
The present invention is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems) and computer program products according to embodiments of the invention. It will be understood that each flow and/or block of the flowchart illustrations and/or block diagrams, and combinations of flows and/or blocks in the flowchart illustrations and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
The foregoing description of the embodiments has been provided for the purpose of illustrating the general principles of the invention, and is not meant to limit the scope of the invention, but to limit the invention to the particular embodiments, and any modifications, equivalents, improvements, etc. that fall within the spirit and principles of the invention are intended to be included within the scope of the invention.
Claims (6)
1. A method for detecting continuous equidistant friction damage of steel rails is characterized by comprising the following steps:
acquiring axle box acceleration signals acquired at different positions by a detection vehicle running on a steel rail to obtain a first signal sequence;
identifying and filtering the welding joint signals in the first signal sequence to obtain a second signal sequence;
performing resonance demodulation processing on the second signal sequence to obtain a third signal sequence;
performing normalization processing on the third signal sequence to obtain the scratch indexes of the different positions;
identifying the scratch positions in the steel rail according to preset conditions; the preset condition comprises that the number of continuous peaks exceeding a first threshold exceeds a second threshold in the scratch index, and the difference between every two peak intervals is smaller than a preset error distance, wherein each peak interval is the distance between two adjacent continuous peaks;
performing normalization processing on the third signal sequence to obtain the scratch indexes of the different positions, including:
normalizing the third signal sequence to be within the numerical range of [0,1] by utilizing a neural network threshold function sigmoid function to obtain the scratch index; wherein normalizing the third signal sequence to within a value range of [0,1] comprises setting an intermediate value of the scratch index to 1/2 by the following formula:
c＝(x max +x min )/2；
Wherein i=1, 2, … N, N is the number of signals of the third signal sequence, W BI,i The scratch index, X of the position corresponding to the ith signal i For the ith signal in the third signal sequence, c is the reference value of the signal in the third signal sequence, x max X is the maximum value in the third signal sequence min A is a first coefficient and a > 0, α is a second coefficient, which is the minimum value in the third signal sequence;
identifying the scratch position in the steel rail according to preset conditions, including:
extracting a local maximum value in the scratch index to obtain a peak value;
calculating the peak distance between every two adjacent peaks;
dividing the scratch index into a plurality of sections with intersections according to a preset track length; wherein the second threshold number of peaks are included in each segment;
searching the sections with the same peak intervals in the sections;
determining that a scratch exists in the corresponding section under the condition that all peaks in the searched section exceed the first threshold value;
identifying and filtering the welding joint signal in the first signal sequence to obtain a second signal sequence, including:
calculating the moving effective value of each signal in the first signal sequence by the following formula to obtain { RMS } i ,i＝1,2,…N}：
The length of the moving window of the moving effective value is K signals;
segmenting the mobile effective value to obtain a plurality of segments;
according to the average value of each segmentAnd variance sigma, a reference threshold R for each segment is calculated by the following formula T ：
Find out that exceeds the reference threshold R T Is recorded as a first set of mobile effective valuesWherein N is R Is->Is the number of (3);
aggregating the first set, and keeping the maximum value in the movement effective values with the distance smaller than the preset distance to obtain a second setWherein (1)>Is->Is the number of (3);
determining the position corresponding to each data in the second set to obtain the position of the welding joint;
supplementing missing welding joint positions in the second set according to the design spacing of the welding joints;
determining a signal corresponding to the welding joint position in the first signal sequence to obtain the welding joint signal;
and setting K signals near the welding joint signals as preset values.
2. The method of claim 1, wherein performing a resonance demodulation process on the second signal sequence to obtain a third signal sequence comprises:
and calculating a Hilbert envelope signal of the second signal sequence through Hilbert transformation to obtain the third signal sequence.
3. A detection device for continuous equidistant rubbing damage of steel rail is characterized by comprising:
the acquisition unit is used for acquiring axle box acceleration signals acquired by a detection vehicle running on a steel rail at different positions to obtain a first signal sequence;
the identification filtering unit is used for identifying and filtering the welding joint signals in the first signal sequence to obtain a second signal sequence;
the resonance demodulation unit is used for performing resonance demodulation processing on the second signal sequence to obtain a third signal sequence;
the normalization processing unit is used for performing normalization processing on the third signal sequence to obtain the scratch indexes of the different positions;
the identifying unit is used for identifying the scratch position in the steel rail according to preset conditions; the preset condition comprises that the number of continuous peaks exceeding a first threshold exceeds a second threshold in the scratch index, and the difference between every two peak intervals is smaller than a preset error distance, wherein each peak interval is the distance between two adjacent continuous peaks;
the normalization processing unit is further used for normalizing the third signal sequence to be in the numerical range of [0,1] by utilizing a neural network threshold function sigmoid function to obtain the scratch index;
The normalization processing unit is further configured to set the intermediate value of the scratch index to 1/2 by the following formula:
c＝(x max +x min )/2；
wherein i=1, 2, … N, N is the number of signals of the third signal sequence, W BI,i The scratch index, X of the position corresponding to the ith signal i For the ith signal in the third signal sequence, c is the reference value of the signal in the third signal sequence, x max X is the maximum value in the third signal sequence min A is a first coefficient and a > 0, α is a second coefficient, which is the minimum value in the third signal sequence;
the identification unit includes:
the extraction unit is used for extracting a local maximum value in the punch damage index to obtain a peak value;
a first calculation unit for calculating a peak interval between every two adjacent peaks;
a dividing unit for dividing the scratch index into a plurality of sections having intersections according to a preset track length; wherein the second threshold number of peaks are included in each segment;
the searching unit is used for searching the sections with the same peak intervals in the sections;
a first determining unit, configured to determine that a scratch exists in a corresponding section if all peaks in the searched section exceed the first threshold;
The identification filtering unit comprises:
a second calculation unit for calculating the movement effective value of each signal in the first signal sequence by the following formula to obtain { RMS } i ,i＝1,2,…N}：
The length of the moving window of the moving effective value is K signals;
the segmentation unit is used for segmenting the mobile effective value to obtain a plurality of segments;
a third calculation unit for calculating an average value of each segmentAnd variance sigma, a reference threshold R for each segment is calculated by the following formula T ：
A searching unit for searching for the reference threshold R T Is recorded as a first set of mobile effective valuesWherein N is R Is->Is the number of (3);
an aggregation unit for aggregating the first set, and keeping the maximum value in the movement effective value smaller than the preset distance to obtain a second setWherein (1)>Is->Is the number of (3);
the second determining unit is used for determining the position corresponding to each data in the second set to obtain the position of the welding joint;
the supplementing unit is used for supplementing the missing welding joint positions in the second set according to the design interval of the welding joints;
a third determining unit, configured to determine a signal corresponding to the welding joint position in the first signal sequence, so as to obtain the welding joint signal;
And the setting unit is used for setting the welding joint signal and the K signals nearby as preset values.
4. The apparatus of claim 3, wherein the resonance demodulation unit is further configured to calculate a hilbert envelope signal of the second signal sequence by a hilbert transform to obtain the third signal sequence.
5. A computer device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, characterized in that the processor implements the method of any of claims 1 to 2 when executing the computer program.
6. A computer readable storage medium, characterized in that the computer readable storage medium stores a computer program which, when executed by a processor, implements the method of any of claims 1 to 2.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110744451.6A CN113343919B (en) | 2021-06-30 | 2021-06-30 | Method and device for detecting continuous equidistant rubbing damage of steel rail and computer equipment |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110744451.6A CN113343919B (en) | 2021-06-30 | 2021-06-30 | Method and device for detecting continuous equidistant rubbing damage of steel rail and computer equipment |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN113343919A CN113343919A (en) | 2021-09-03 |
| CN113343919B true CN113343919B (en) | 2024-03-15 |
Family
ID=77482072
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110744451.6A Active CN113343919B (en) | 2021-06-30 | 2021-06-30 | Method and device for detecting continuous equidistant rubbing damage of steel rail and computer equipment |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113343919B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113830132B (en) * | 2021-09-27 | 2022-10-25 | 中国铁道科学研究院集团有限公司 | Method and device for detecting arching of track slab |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2002107348A (en) * | 2000-09-28 | 2002-04-10 | Meitetsu Sumisho Kogyo Kk | Method and device for detecting damaged state of wheel tread |
| JP2004205403A (en) * | 2002-12-26 | 2004-07-22 | Meitetsu Sumisho Kogyo Kk | Damage state detection method and device of wheel tread |
| CN105510438A (en) * | 2015-12-02 | 2016-04-20 | 北京交通大学 | Operating vehicle-based rail nucleus flaw detection system and method |
| CN107244335A (en) * | 2017-07-20 | 2017-10-13 | 中国铁道科学研究院基础设施检测研究所 | A kind of undulatory wear detection means and method |
| CN108845028A (en) * | 2018-03-26 | 2018-11-20 | 中国铁路总公司 | A kind of rail in high speed railway wave mill dynamic testing method and device |
| WO2020258106A1 (en) * | 2019-06-26 | 2020-12-30 | Oppo广东移动通信有限公司 | Gesture recognition method and device, and positioning and tracking method and device |
-
2021
- 2021-06-30 CN CN202110744451.6A patent/CN113343919B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2002107348A (en) * | 2000-09-28 | 2002-04-10 | Meitetsu Sumisho Kogyo Kk | Method and device for detecting damaged state of wheel tread |
| JP2004205403A (en) * | 2002-12-26 | 2004-07-22 | Meitetsu Sumisho Kogyo Kk | Damage state detection method and device of wheel tread |
| CN105510438A (en) * | 2015-12-02 | 2016-04-20 | 北京交通大学 | Operating vehicle-based rail nucleus flaw detection system and method |
| CN107244335A (en) * | 2017-07-20 | 2017-10-13 | 中国铁道科学研究院基础设施检测研究所 | A kind of undulatory wear detection means and method |
| CN108845028A (en) * | 2018-03-26 | 2018-11-20 | 中国铁路总公司 | A kind of rail in high speed railway wave mill dynamic testing method and device |
| WO2020258106A1 (en) * | 2019-06-26 | 2020-12-30 | Oppo广东移动通信有限公司 | Gesture recognition method and device, and positioning and tracking method and device |
Non-Patent Citations (2)
| Title |
|---|
| 列车车轮踏面擦伤信号处理算法研究;马莉;傅八路;王泽勇;王黎;杨凯;;铁道技术监督(03);全文 * |
| 基于连续子波变换的铁路车轮踏面擦伤的在线检测;岳建海,裘正定,李铁锚;铁道学报(04);全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN113343919A (en) | 2021-09-03 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108845028B (en) | Method and device for dynamically detecting high-speed railway rail corrugation | |
| Molodova et al. | Automatic detection of squats in railway infrastructure | |
| CN105092467B (en) | A kind of quick monitoring device of high-speed railway wheel tread flat and method | |
| CN109030628B (en) | Track structure mechanical property evaluation method based on vibration noise frequency spectrum characteristic matching | |
| CN110987348B (en) | Catenary hard spot determination method and device based on pantograph-catenary dynamic response | |
| CN108515984B (en) | Wheel damage detection method and device | |
| CN106018557B (en) | A kind of raceway surface recess detection method based on wavelet analysis | |
| CN105929025B (en) | One kind is based on the continuous wheel tread of the time and space and rail deformation detection method | |
| CN113343928B (en) | Method and device for detecting rail wave mill of speed change section high-speed railway and computer equipment | |
| CN113343919B (en) | Method and device for detecting continuous equidistant rubbing damage of steel rail and computer equipment | |
| CN113947130B (en) | Training and using method and device for regression prediction AI model of polygonal wear waveform of wheel | |
| CN106218668A (en) | Wheel out of round degree detection method and device | |
| CN107521511B (en) | A kind of method and system of wheel derailment alarm | |
| DE102019210884B4 (en) | Measuring arrangement and method for determining a distance between a distance sensor arranged on an axle-mounted gearbox of a rail-bound vehicle and a measuring zero surface outside the gearbox during operation | |
| CN104457644B (en) | The detection method and device of non-pulse abnormal data in track geometry detection data | |
| JP7049198B2 (en) | Wheel uneven wear judgment method and judgment program | |
| CN107782548B (en) | Rail vehicle part detection system | |
| CN116252820B (en) | Polygonal quantitative detection method for high-speed train wheels driven by improved frequency domain integration method | |
| CN113032907A (en) | Method and system for correcting vehicle shaking disease data deviation based on waveform correlation | |
| CN115931399A (en) | Method for real-time online detection of polygonal fault of high-speed train wheel | |
| CN108827665A (en) | Wheel flat fault detection method based on empirical mode decomposition and multi-scale entropy | |
| CN114348054B (en) | Railway switch disease monitoring method, device and system | |
| CN114580460A (en) | Railway vehicle wheel rail fault diagnosis method based on morphological filtering and HHT conversion | |
| CN109211556B (en) | Rail vehicle spare part detecting system | |
| JP2008120258A (en) | Detection system of wheel tread state |
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 |