CN107368779B - Feature extraction method of turnout action current curve - Google Patents

Feature extraction method of turnout action current curve Download PDF

Info

Publication number
CN107368779B
CN107368779B CN201710409424.7A CN201710409424A CN107368779B CN 107368779 B CN107368779 B CN 107368779B CN 201710409424 A CN201710409424 A CN 201710409424A CN 107368779 B CN107368779 B CN 107368779B
Authority
CN
China
Prior art keywords
variance
turnout
current
curve
current data
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
Application number
CN201710409424.7A
Other languages
Chinese (zh)
Other versions
CN107368779A (en
Inventor
赵春阳
王彬
张硕
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Henan Splendor Science and Technology Co Ltd
Original Assignee
Henan Splendor Science and Technology Co Ltd
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Henan Splendor Science and Technology Co Ltd filed Critical Henan Splendor Science and Technology Co Ltd
Priority to CN201710409424.7A priority Critical patent/CN107368779B/en
Publication of CN107368779A publication Critical patent/CN107368779A/en
Application granted granted Critical
Publication of CN107368779B publication Critical patent/CN107368779B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
    • G06F2218/08—Feature extraction
    • B—PERFORMING OPERATIONS; TRANSPORTING
    • B61—RAILWAYS
    • B61L—GUIDING RAILWAY TRAFFIC; ENSURING THE SAFETY OF RAILWAY TRAFFIC
    • B61L23/00—Control, warning or like safety means along the route or between vehicles or trains
    • B61L23/04—Control, warning or like safety means along the route or between vehicles or trains for monitoring the mechanical state of the route

Landscapes

  • Engineering & Computer Science (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Theoretical Computer Science (AREA)
  • Artificial Intelligence (AREA)
  • Signal Processing (AREA)
  • Physics & Mathematics (AREA)
  • General Engineering & Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Mechanical Engineering (AREA)
  • Train Traffic Observation, Control, And Security (AREA)

Abstract

The invention provides a turnout motionThe characteristic extraction method for making the current curve comprises the following steps: sampling turnout action current, and establishing a turnout action current curve; marking the last current data in the turnout action current curve as a conversion end point D; finding out the current data with the maximum current amplitude as a turnout starting point A; sequentially calculating the average value of the jth current data point between the turnout starting point A and the conversion end point D and the n-1 current data points after the jth current data point by a sliding average method
Figure DDA0001311930370000011
And variance KjThe variance KjComparing with a preset threshold value x when K isjWhen x is less than or equal to x, dividing the variance KjHas a value of 1 when Kj>x, the variance KjIs set to 0; establishing a variance curve chart according to the variance KjAnd variance Kj‑1Variance Kj+1The current data corresponding to the starting point of the first rising edge in the variance graph is marked as the starting point B of the turnout action current, and the current data corresponding to the end point of the last falling edge is marked as the end point C of the turnout action current.

Description

Feature extraction method of turnout action current curve
Technical Field
The invention relates to a turnout fault diagnosis method in the railway field, in particular to a characteristic extraction method of a turnout action current curve.
Background
The main function of the turnout is to guide the running direction of wheels to realize the line-changing and line-crossing running of the train, and the turnout is one of key devices for guaranteeing the safety and efficiency of railway transportation. The turnout action current curve is an important index reflecting the turnout operation quality, and is a standard graph of the turnout action current curve as shown in fig. 1, and the standard graph displays the action time, direction, duration and current curve waveform in the turnout action process. Because the turnout action curve is influenced by various environments and has large individual difference, the turnout action curve lacks a quantitative analysis method, signal maintainers often need to judge the turnout action process through personal experience, and the abnormal condition in the turnout action process lacks automatic analysis.
In order to solve the above problems, people are always seeking an ideal technical solution.
Disclosure of Invention
The invention aims to overcome the defects of the prior art and provides a method for extracting characteristics of a turnout action current curve.
In order to achieve the purpose, the invention adopts the technical scheme that: a feature extraction method of a switch action current curve comprises the following steps:
step 1, sampling turnout action current, acquiring current data of m turnout actions, and establishing a turnout action current curve by taking turnout action time as an abscissa and discrete current data as an ordinate;
step 2, marking the mth current data in the turnout action current curve as a conversion end point D;
step 3, traversing all current data in the turnout action current curve, and finding out the current data with the maximum current amplitude as a turnout starting point A;
step 4, in the turnout action current curve, sequentially calculating the jth current data point between the turnout starting point A and the conversion end point D and the average value of the subsequent n-1 current data points by a sliding average method
Figure BDA0001311930350000021
And variance KjThe a-th current data is a turnout starting point A, j belongs to (a, m);
step 5, the variance K is calculatedjComparing with a preset threshold value x when K isjSetting the variance value to 1 when x is less than or equal to the value, and setting the variance value to 1 when K is less than or equal to xj>When x is needed, setting the value of the variance as 0;
step 6, establishing a variance curve chart and according to the variance KjAnd variance Kj-1Variance Kj+1And (3) searching all rising edge starting points and all falling edge end points in the variance graph, marking the current data corresponding to the first rising edge starting point as a turnout action current starting point B, and marking the current data corresponding to the last falling edge end point as a turnout action current end point C.
Based on the above, in step 6, all the variance graphs in the variance graph are searchedThe specific steps of the starting point of the rising edge and the end points of all the falling edges are as follows: go through j, and sum the variance KjAnd its pre-variance Kj-1Its posterior variance Kj+1Making a comparison while averaging
Figure BDA0001311930350000022
And its pre-average value
Figure BDA0001311930350000023
Mean value of the mean value
Figure BDA0001311930350000024
Making a comparison when Kj-1＜Kj＝Kj+1And is
Figure BDA0001311930350000025
Then, the variance K is recordedjIs the starting point of the rising edge of the variance curve; when K isj+1＜Kj＝Kj-1And is
Figure BDA0001311930350000026
Then, the variance K is recordedjThe end of the falling edge of the variance curve.
Compared with the prior art, the turnout current action curve analysis method has outstanding substantive characteristics and obvious progress, particularly, a normal turnout current action curve is analyzed by three stages of unlocking, conversion and locking, the key characteristic value of the turnout current curve is extracted by a mathematical calculation method according to the sectional analysis characteristics of the turnout current action curve and the characteristics of 4 end points corresponding to the single stage, the turnout conversion stage division and the subsequent fault mode identification work are greatly simplified, the calculation complexity of fault diagnosis is greatly reduced, the diagnosis accuracy is greatly improved finally, and a solid foundation is laid for improving the efficiency of turnout monitoring and maintenance work.
Drawings
Fig. 1 is a standard diagram of a turnout operating current curve.
Fig. 2 is a flow chart of the present invention.
Fig. 3 is an example of a switch action current curve.
Detailed Description
The technical solution of the present invention is further described in detail by the following embodiments.
As shown in fig. 2, a method for extracting characteristics of a switch action current curve includes the following steps:
step 1, sampling turnout action currents at equal intervals, acquiring current data of m turnout actions, and establishing a turnout action current curve by taking turnout action time as an abscissa and discrete current data as an ordinate;
step 2, marking the mth current data in the turnout action current curve as a conversion end point D, and marking the abscissa of the conversion end point D as turnout conversion end time T4;
step 3, traversing all current data in the turnout action current curve, finding out that the current data with the maximum current amplitude is marked as a turnout starting point A, and marking the abscissa of the turnout starting point A as turnout starting time T1;
step 4, in the turnout action current curve, sequentially calculating the jth current data point between the turnout starting point A and the conversion end point D and the average value of the subsequent n-1 current data points by a sliding average method
Figure BDA0001311930350000035
And variance KjThe a-th current data is a turnout starting point A, j belongs to (a, m);
step 5, the variance K is calculatedjComparing with a preset threshold value x when K isjSetting the variance value to 1 when x is less than or equal to the value, and setting the variance value to 1 when K is less than or equal to xj>When x is needed, setting the value of the variance as 0;
step 6, establishing a variance curve chart and according to the variance KjAnd variance Kj-1Variance Kj+1The relationship (c) is obtained by searching all rising edge starting points and all falling edge end points in the variance graph, marking the current data corresponding to the first rising edge starting point as a turnout action current starting point B, marking the abscissa of the turnout action current starting point B as turnout action starting time T2, and aligning the last falling edge end pointThe current data to be measured is designated as a switch operation current end point C, and the abscissa of the switch operation current end point C is designated as a switch operation end time T3.
Wherein, T4 is the turnout ending time, T1 is the turnout starting time, the difference between T2 and T3 is the turnout operation time, and the difference between T3 and T4 is the turnout locking time.
Specifically, in step 6, the specific steps of searching all rising edge starting points and all falling edge ending points in the variance graph are as follows: go through j, and sum the variance KjAnd its pre-variance Kj-1Its posterior variance Kj+1Making a comparison while averaging
Figure BDA0001311930350000031
And its pre-average value
Figure BDA0001311930350000032
Mean value of the mean value
Figure BDA0001311930350000033
Making a comparison when Kj-1＜Kj＝Kj+1And is
Figure BDA0001311930350000034
Then, the variance K is recordedjIs the starting point of the rising edge of the variance curve; when K isj+1＜Kj＝Kj-1And is
Figure BDA0001311930350000041
Then, the variance K is recordedjThe end of the falling edge of the variance curve.
In order to facilitate clear and intuitive understanding of the present invention, a standard diagram of a switch action current curve as shown in fig. 1 is taken as an example:
sampling turnout operating current at equal intervals of 50ms, and taking the suction and the falling of a relay 1DQJ as the starting and ending points of data acquisition each time to obtain 130 AD data, wherein the 130 AD data are specifically as follows:
Figure BDA0001311930350000042
for identification, the AD data in the 16-ary number format is further converted into a current value through an AD-current conversion formula:
current (AD × (ADMax-ADMin)/ADn + ADMin) × Cx + Cp
Wherein, ADMax is the maximum value of measurement: 10A; ADMin is the minimum measured: 0A; ADn is AD number: 8 bit AD, 256; cx is the linear coefficient: 1.0; cp is the cheap coefficient: 0;
therefore, the simplified formula is: current AD 10/256.
Searching current data with the maximum current amplitude, namely current data corresponding to data with an AD value of 4A, as a turnout starting point A by adopting a bubbling sequencing method; the 130 th current data in the turnout action current curve is marked as a conversion end point D.
Calculating the starting point A and the end of the conversion in the turnout action current curve by a moving average method
Average of each current data point between points D and 4 current data points thereafter
Figure BDA0001311930350000043
Figure BDA0001311930350000044
Wherein, the 6 th current data is the starting points A and T of the turnoutiThe current value of the jth current data point;
calculating the variance K of each current data point between the starting point A and the end point D of the switch and 4 current data points after the current data point according to the variance formulaj：
Figure BDA0001311930350000051
The variance KjComparing with a preset threshold value x, preferably, the preset threshold value x is 3; when K isjSetting the variance value to 1 when the variance value is less than or equal to 3, and setting the variance value to be 1 when K is less than or equal toj>When 3, setting the value of the variance as 0;
go through j, and sum the variance KjValue and its pre-variance Kj-1Its posterior variance Kj+1Making a comparison while averaging
Figure BDA0001311930350000052
And its pre-average value
Figure BDA0001311930350000053
Mean value of the mean value
Figure BDA0001311930350000054
Making a comparison when Kj-1＝0,Kj＝K j+11 and
Figure BDA0001311930350000055
when, illustrate the variance KjIs the starting point of the rising edge of the variance curve; when K isj-1＝1,Kj＝Kj+1Is equal to 0 and
Figure BDA0001311930350000056
when it is, the variance K is describedjThe end of the falling edge of the variance curve.
The data between each rising edge starting point and the subsequent falling edge ending point is the upper step current or the upper peak current, the data between each falling edge ending point and the subsequent rising edge starting point is the lower step current or the lower peak current, and the data between the last falling edge ending point and the data tail is the locking current; the number O of the upper and lower peaks and the step current after filtering reasonable fluctuation can be calculated through the steps.
And finally, marking the current data corresponding to the starting point of the first rising edge as a turnout action current starting point B, and marking the current data corresponding to the end point of the last falling edge as a turnout action current end point C, namely finishing the extraction of all key points and key data, and further performing subsection analysis on the waveform of the action curve of the switch machine.
Wherein, the starting current peak value is the vertical coordinate of the point A;
action current smoothness: average value after the waveform of the burrs between the point B and the point C;
the turnout action current: average value after filtering burr waveform between B point and C point;
and in practical application, the current curve of each action of the turnout is compared with the key data of the standard curve of the turnout action current, so that whether the turnout equipment is in a good state or not is judged. As shown in fig. 3, an example of the switch action current is shown, specifically, 130 AD data are as follows:
Figure BDA0001311930350000061
the turnout current curve shown in FIG. 3 was analyzed by using the above algorithm, and the analysis results are shown in Table 1:
judging the content Example values Standard of merit Overrun
Starting current (A) 2.89 5 0
Conversion time (S) 6.50 6.8 0
Starting time (S) 0.30 0.5 0
Action current (A) 0.87 2 0
Conversion time (S) 4.70 5 0
Fluctuation of curve 5 3 1
Locking time (S) 1.05 1.5 0
TABLE 1
As can be seen from Table 1, the number of current fluctuation times of the turnout operation in the example is 5, which is greater than the number of current fluctuation times 3 of the standard diagram, and the turnout has electrical or mechanical problems.
Finally, it should be noted that the above examples are only used to illustrate the technical solutions of the present invention and not to limit the same; although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art will understand that: modifications to the specific embodiments of the invention or equivalent substitutions for parts of the technical features may be made; without departing from the spirit of the present invention, it is intended to cover all aspects of the invention as defined by the appended claims.

Claims (2)

1. A feature extraction method of a switch action current curve is characterized by comprising the following steps:
step 1, sampling turnout action current, acquiring current data of m turnout actions, and establishing a turnout action current curve by taking turnout action time as an abscissa and discrete current data as an ordinate;
step 2, marking the mth current data in the turnout action current curve as a conversion end point D;
step 3, traversing all current data in the turnout action current curve, and finding out the current data with the maximum current amplitude as a turnout starting point A;
step 4, in the turnout action current curve, sequentially calculating the jth current data point between the turnout starting point A and the conversion end point D and the average value of the subsequent n-1 current data points by a sliding average method
Figure FDA0001311930340000016
And variance KjThe a-th current data is a turnout starting point A, j belongs to (a, m);
step 5, the variance K is calculatedjComparing with a preset threshold value x when K isjWhen x is less than or equal to x, dividing the variance KjHas a value of 1 when Kj>x, the variance KjIs set to 0;
step 6, establishing a variance curve chart and according to the variance KjAnd variance Kj-1Variance Kj+1And (3) searching all rising edge starting points and all falling edge end points in the variance graph, marking the current data corresponding to the first rising edge starting point as a turnout action current starting point B, and marking the current data corresponding to the last falling edge end point as a turnout action current end point C.
2. The method for extracting characteristics of a turnout action current curve according to claim 1, wherein in step 6, the specific steps of searching all rising edge starting points and all falling edge end points in the variance graph are as follows: go through j, and sum the variance KjAnd its pre-variance Kj-1After thatVariance Kj+1Making a comparison while averaging
Figure FDA0001311930340000011
And its pre-average value
Figure FDA0001311930340000012
Mean value of the mean value
Figure FDA0001311930340000013
Making a comparison when Kj-1＜Kj＝Kj+1And is
Figure FDA0001311930340000014
Then, the variance K is recordedjIs the starting point of the rising edge of the variance curve; when K isj+1＜Kj＝Kj-1And is
Figure FDA0001311930340000015
Then, the variance K is recordedjThe end of the falling edge of the variance curve.
CN201710409424.7A 2017-06-02 2017-06-02 Feature extraction method of turnout action current curve Active CN107368779B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201710409424.7A CN107368779B (en) 2017-06-02 2017-06-02 Feature extraction method of turnout action current curve

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201710409424.7A CN107368779B (en) 2017-06-02 2017-06-02 Feature extraction method of turnout action current curve

Publications (2)

Publication Number Publication Date
CN107368779A CN107368779A (en) 2017-11-21
CN107368779B true CN107368779B (en) 2020-05-29

Family

ID=60306459

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201710409424.7A Active CN107368779B (en) 2017-06-02 2017-06-02 Feature extraction method of turnout action current curve

Country Status (1)

Country Link
CN (1) CN107368779B (en)

Families Citing this family (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN108287273B (en) * 2017-12-13 2020-07-24 河南辉煌科技股份有限公司 Current curve-based speed-up turnout jamming fault judgment method
CN110503069B (en) * 2019-08-28 2022-10-18 中广核研究院有限公司 Current waveform fluctuation starting point identification method, electronic device and readable storage medium
CN112084673B (en) * 2020-09-17 2023-01-31 广西交控智维科技发展有限公司 Automatic setting method and device for switch friction current
CN117033918B (en) * 2023-08-01 2024-06-11 珠海精实测控技术股份有限公司 Waveform data segmentation processing method and device

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101513884A (en) * 2008-12-29 2009-08-26 卡斯柯信号有限公司 Method for analyzing and processing turnout curves
CN105260595A (en) * 2015-04-02 2016-01-20 北京交通大学 Feature extraction method for switch action current curve and switch fault diagnosis method
CN106124885A (en) * 2016-06-13 2016-11-16 四川网达科技有限公司 Switch breakdown detection apparatus and method
CN106184288A (en) * 2016-08-15 2016-12-07 绵阳市维博电子有限责任公司 A kind of direct current track switch notch state detection method

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101513884A (en) * 2008-12-29 2009-08-26 卡斯柯信号有限公司 Method for analyzing and processing turnout curves
CN105260595A (en) * 2015-04-02 2016-01-20 北京交通大学 Feature extraction method for switch action current curve and switch fault diagnosis method
CN106124885A (en) * 2016-06-13 2016-11-16 四川网达科技有限公司 Switch breakdown detection apparatus and method
CN106184288A (en) * 2016-08-15 2016-12-07 绵阳市维博电子有限责任公司 A kind of direct current track switch notch state detection method

Also Published As

Publication number Publication date
CN107368779A (en) 2017-11-21

Similar Documents

Publication Publication Date Title
CN107368779B (en) Feature extraction method of turnout action current curve
CN108416362B (en) Turnout abnormity early warning and fault diagnosis method
CN101919695B (en) Electrocardiosignal QRS complex detection method based on wavelet transform
CN107451004B (en) Turnout fault diagnosis method based on qualitative trend analysis
KR101538843B1 (en) Yield management system and method for root cause analysis using manufacturing sensor data
CN109828184B (en) Voltage sag source identification method based on mutual approximate entropy
EP3764361B1 (en) Method and apparatus for aligning voices
CN104545887A (en) Method and device for identifying artifact electrocardiograph waveforms
CN105030233A (en) Method for recognizing ST segment of electrocardiosignal
CN106740990A (en) Track switch operating power Curves Recognition method and system
CN104849645A (en) MOSFET degeneration assessment method based on Miller platform voltage, and MOSFET residual life prediction method applying the method
CN116739829B (en) Big data-based power data analysis method, system and medium
CN110658445A (en) Analysis and diagnosis method for mechanical fault of on-load tap-changer
CN113567127B (en) Rolling bearing degradation index extraction method based on time-frequency feature separation
CN112116013A (en) Voltage sag event normalization method based on waveform characteristics
CN109359672A (en) A kind of oil level gauge for transformer reading image-recognizing method
CN108594156B (en) Improved current transformer saturation characteristic identification method
CN109490776B (en) Mobile phone vibration motor good and defective product detection method based on machine learning
CN109596354B (en) Band-pass filtering method based on self-adaptive resonance frequency band identification
CN108173792B (en) Wireless device transient characteristic extraction and identification method based on differential constellation locus diagram
CN108153711A (en) A kind of electrical equipment online supervision data processing method
CN110490297B (en) Intelligent segmentation method for railway turnout action power curve
CN116317103A (en) Power distribution network voltage data processing method
CN116626454B (en) Oil paper insulation UHF partial discharge signal anti-interference identification and positioning method and device based on correction time-frequency clustering
CN110991363A (en) Method for extracting CO emission characteristics of coal mine safety monitoring system under different coal mining processes

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