CN106323442B - A kind of railway health monitor method based on distributed optical fiber vibration sensing system - Google Patents
A kind of railway health monitor method based on distributed optical fiber vibration sensing system Download PDFInfo
- Publication number
- CN106323442B CN106323442B CN201610688251.2A CN201610688251A CN106323442B CN 106323442 B CN106323442 B CN 106323442B CN 201610688251 A CN201610688251 A CN 201610688251A CN 106323442 B CN106323442 B CN 106323442B
- Authority
- CN
- China
- Prior art keywords
- rail
- characteristic frequency
- vibration
- wheel
- optical fiber
- 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
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01H—MEASUREMENT OF MECHANICAL VIBRATIONS OR ULTRASONIC, SONIC OR INFRASONIC WAVES
- G01H9/00—Measuring mechanical vibrations or ultrasonic, sonic or infrasonic waves by using radiation-sensitive means, e.g. optical means
- G01H9/004—Measuring mechanical vibrations or ultrasonic, sonic or infrasonic waves by using radiation-sensitive means, e.g. optical means using fibre optic sensors
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01M—TESTING STATIC OR DYNAMIC BALANCE OF MACHINES OR STRUCTURES; TESTING OF STRUCTURES OR APPARATUS, NOT OTHERWISE PROVIDED FOR
- G01M17/00—Testing of vehicles
- G01M17/08—Railway vehicles
Abstract
The invention discloses a kind of high-speed railway health monitor method based on distributed optical fiber vibration sensing system, the existing telecommunication optical fiber being laid with using Along Railway are sensed, vibration information of the capture Along Railway after train passes through.By data, countershaft is organized as the Waterfall plot that a gray scale indicates intensity on time, and after carrying out wavelet decomposition filtering, the track of train driving generation is extracted using the Boundary extracting algorithm based on Dynamic Programming, marks off before track and latter two region of track.Time frequency analysis is carried out to region before and after track regions and track respectively, Wheel Rail Vibration characteristic frequency spectrum and rail resonance-characteristic frequency spectrum is extracted, using characteristic frequency spectrum extraction algorithm, obtains Wheel Rail Vibration characteristic frequency and rail resonance-characteristic frequency.This monitoring process is constantly repeated, rail vibration characteristic is obtained and changes with time, establish corresponding database.New collected data are compared and are updated the data library with the data in database, are realized and are monitored to the real time health of train and rail.
Description
Technical field
The present invention relates to technical field of optical fiber sensing, especially a kind of railway based on distributed optical fiber vibration sensing system
Health monitor method.
Background technology
Railway operation safety is always to be related to the significant problem of economic development and personal safety.In order to ensure that railway is transported
Defeated safety with it is unimpeded, improve efficiency of operation, railway health status is monitored in real time very necessary.However use at present
The moisture-proof wetting resistance of electric class sensing element and anti-electromagnetic interference capability are poor, showed under the conditions of severe monitoring it is unstable, when
When being used under complicated monitoring of environmental for a long time, be easy to happen the failures such as null offset, strong influence monitoring result it is reliable
Property；On the other hand, the transmission range of electric signal in the channel is very short, it is difficult to which group builds up large-scale sensing network, it is difficult to realize
Long range real time on-line monitoring.
Distributed optical fiber vibration sensing system as a kind of novel security protection monitoring system, not only have electromagnetism interference,
The features such as anticorrosive, high sensitivity, and have many advantages, such as that good concealment, alarm registration, data processing are relatively easy, it fits
It shares in a wide range of, monitoring in real time over long distances.
A kind of preferred embodiment by phase sensitive optical time domain reflectometer (Φ-OTDR) as optical fiber vibration sensing system, tool
There is fast response time and can realize the clear advantages such as multiple spot monitoring.Compared to Brillouin scattering and Raman scattering light measurement
Vibration, Φ-OTDR are not necessarily to multiple cumulative mean, thus fast response time；Compared with polarization-optical time domain reflectometry (POTDR),
Very noisy bottom not will produce at train rear using Φ-OTDR, multiple spot may be implemented and monitor in real time.
From in china railway main line, along with the laying of rail, being laid with mass communication optical fiber along the line since 2000,
It has been basically formed to today by the communication network that optical fiber is transmission medium.The existing Communication ray being laid with using Along Railway
Fibre, without relying on special laying sensor fibre, can save a large amount of costs for purchasing sensor fibre can group convenient for construction
Build up large-scale sensing network.But telecommunication optical fiber, which is often placed on, to be had with rail in the cable trough of certain distance, not directly
Rail and roadbed are contacted, keeps certain distance with rail, less efficient on vibration coupling to optical fiber, this causes to sensitivity
It is required that harsher.
Invention content
It is provided the technical problem to be solved by the present invention is to overcome the deficiencies in the prior art a kind of based on distribution type fiber-optic
The railway health monitor method of vibration sensing system, the present invention are laid with Along Railway using distributed optical fiber vibration sensing system
Telecommunication optical fiber be detected, obtain a large amount of fiber-optic vibration curve data, carrying out analyzing processing to data establishes corresponding number
According to library；The processing data newly measured are compared and analyzed with data in database, progress healthy and safe to railway may be implemented in fact
When monitor.
The present invention uses following technical scheme to solve above-mentioned technical problem：
According to a kind of railway health monitor method based on distributed optical fiber vibration sensing system proposed by the present invention, including
Following steps：
Step 1: being sensed using the existing telecommunication optical fiber that Along Railway is laid with, capture Along Railway is respectively in train
Through vibration information later along the Two dimensional Distribution situation in time and space；On time by measured obtained fiber-optic vibration curve data
Countershaft is organized as the Waterfall plot that a gray scale indicates intensity, and the abscissa of the Waterfall plot is fiber lengths information, ordinate is the time
Length information；
Step 2: to there is the Waterfall plot corresponding to the period that train crosses to be filtered, using based on Dynamic Programming
Boundary extracting algorithm extraction train crosses the trajectory diagram of tested rail, includes Wheel Rail Vibration relation data in trajectory diagram；In track
Former and later two ranges of track are marked off in figure, include rail vibration data within the scope of the two；
Step 3: carrying out time frequency analysis processing respectively to Wheel Rail Vibration relation data and rail vibration data, wheel is extracted
Rail vibration performance frequency spectrum and rail resonance-characteristic frequency spectrum, Wheel Rail Vibration characteristic frequency spectrum and rail resonance-characteristic frequency spectrum are passed through respectively again
Characteristic frequency spectrum extraction algorithm is crossed, Wheel Rail Vibration characteristic frequency and rail resonance-characteristic frequency are obtained；
Step 4: the Waterfall plot corresponding to the period crossed to no train carries out time-domain analysis processing, extraction background is made an uproar
The characteristic frequency spectrum of sound obtains the feature frequency of ambient noise by the characteristic frequency spectrum of ambient noise using characteristic frequency spectrum extraction algorithm
Rate；
Step 5: it is undamaged on the surface of rail and wheel, step 1 is repeated to step 4, establishes wheel track
Relational database and background noise data library；By Wheel Rail Vibration characteristic frequency and rail resonance-characteristic frequency and corresponding train type
Number and rail road section information be stored in wheel rail relation database；The characteristic frequency of ambient noise and rail road section information are stored
In background noise data library；
Step 6: the Wheel Rail Vibration characteristic frequency obtained in step 3 and rail resonance-characteristic frequency are established with step 5
Wheel rail relation database be compared；It is specific as follows：
601, when the difference of Wheel Rail Vibration characteristic frequency and the Wheel Rail Vibration characteristic frequency in wheel rail relation database is pre-
If in allowable range, and rail resonance-characteristic frequency and the gap of the rail resonance-characteristic frequency in wheel rail relation database also exist
In default allowable range；It is then that observation state is stored in wheel rail relation database by this group of data markers；
602, when the difference of the Wheel Rail Vibration characteristic frequency in Wheel Rail Vibration characteristic frequency and wheel rail relation database does not exist
The gap of rail resonance-characteristic frequency in default allowable range or in rail resonance-characteristic frequency and wheel rail relation database
It is that train wheels go wrong if the Wheel Rail Vibration characteristic frequency of whole railway changes when not in default allowable range,
Otherwise, then it is that rail goes wrong, by this group of data markers is that alarm condition is stored in wheel rail relation database, and by train wheels
The information that the information that goes wrong, rail go wrong is transferred to monitoring personnel；
Step 7: data in the characteristic frequency of the ambient noise obtained in step 4 and background noise data library are carried out pair
Than analysis, big data analysis is carried out to background noise data, analyzes situation of change of the rail with external environment, it is old to obtain it
Law and bimetry information.
It is further as a kind of railway health monitor method based on distributed optical fiber vibration sensing system of the present invention
Prioritization scheme, the filtering method in the step 2 are wavelet decomposition denoising.
It is further as a kind of railway health monitor method based on distributed optical fiber vibration sensing system of the present invention
The method of prioritization scheme, time frequency analysis processing is Short Time Fourier Transform or wavelet analysis.
It is further as a kind of railway health monitor method based on distributed optical fiber vibration sensing system of the present invention
Prioritization scheme, characteristic frequency spectrum extraction algorithm are the polynomial fitting method based on least square method.
It is further as a kind of railway health monitor method based on distributed optical fiber vibration sensing system of the present invention
Prioritization scheme senses the existing telecommunication optical fiber that Along Railway is laid with using distributed optical fiber vibration sensing system.
It is further as a kind of railway health monitor method based on distributed optical fiber vibration sensing system of the present invention
Prioritization scheme, distributed optical fiber vibration sensing system are Φ-OTDR systems.
It is further as a kind of railway health monitor method based on distributed optical fiber vibration sensing system of the present invention
Prioritization scheme, the track obtained in analytical procedure two are realized the positioning to train, are measured in the position of spatial axes and time shaft
Speed and acceleration.
The present invention has the following technical effects using above technical scheme is compared with the prior art：
(1) vibration that surrounding is measured using distributed optical fiber sensing system, can realize long range, high density, quickly
The real time health of response monitors, while sensor has good moisture-proof wetting resistance and anti-electricity without being in direct contact with rail
Magnetic disturbance ability shows excellent under the conditions of performance steady in a long-term is in severe monitoring；The present invention uses existing in railway
Telecommunication optical fiber, group are built up large-scale anti-interference strong optical fiber sensing network and are saved, and the cost for arranging new fiber optic network is gone；
(3) distributed optical fiber sensing system, realization is used to position to train, measured to speed and acceleration；
Train and rail can be monitored in real time, go wrong timely discovery, and alarms to train dispatcher, it occurs in accident
Preceding timely reparing security risk ensures railway health detection.
Description of the drawings
Fig. 1 is the flow chart of the present invention.
Fig. 2 is the spatial relation graph of optical fiber and rail of the present invention.
Fig. 3 is that Φ-OTDR measure original signal figure.
Fig. 4 is processing rank rear wheel paths and front and back administrative division map.
Fig. 5 a are wheel track frequency relation figures.
Fig. 5 b are rail characteristic frequency figures.
Fig. 6 is rail characteristic frequency along rail distribution map.
Fig. 7 is characteristic frequency change frequency figure.
Specific implementation mode
Technical scheme of the present invention is described in further detail below in conjunction with the accompanying drawings：
Fig. 1 be the present invention flow chart, a kind of railway health monitor method based on distributed optical fiber vibration sensing system,
Include the following steps：
Step 1: being sensed using the existing telecommunication optical fiber that Along Railway is laid with, capture Along Railway is respectively in train
Through vibration information later along the Two dimensional Distribution situation in time and space.On time by measured obtained fiber-optic vibration curve data
Countershaft is organized as the Waterfall plot that a gray scale indicates intensity, and the abscissa of the Waterfall plot is fiber lengths information, ordinate is the time
Length information.
Step 2: to there is the Waterfall plot corresponding to the period that train crosses to be filtered, using based on Dynamic Programming
Boundary extracting algorithm extraction train crosses the trajectory diagram of tested rail, includes Wheel Rail Vibration relation data in trajectory diagram；In track
Former and later two ranges of track are marked off in figure, include rail vibration data within the scope of the two；
Step 3: carrying out time frequency analysis processing respectively to Wheel Rail Vibration relation data and rail vibration data, wheel is extracted
Rail vibration performance frequency spectrum and rail resonance-characteristic frequency spectrum, Wheel Rail Vibration characteristic frequency spectrum and rail resonance-characteristic frequency spectrum are passed through respectively again
Characteristic frequency spectrum extraction algorithm is crossed, Wheel Rail Vibration characteristic frequency and rail resonance-characteristic frequency are obtained；
Step 4: the Waterfall plot corresponding to the period crossed to no train carries out time-domain analysis processing, extraction background is made an uproar
The characteristic frequency spectrum of sound obtains the feature frequency of ambient noise by the characteristic frequency spectrum of ambient noise using characteristic frequency spectrum extraction algorithm
Rate；
Step 5: it is undamaged on the surface of rail and wheel, step 1 is repeated to step 4, establishes wheel track
Relational database and background noise data library；By Wheel Rail Vibration characteristic frequency and rail resonance-characteristic frequency and corresponding train type
Number and rail road section information be stored in wheel rail relation database；The characteristic frequency of ambient noise and rail road section information are stored
In background noise data library；
Step 6: the Wheel Rail Vibration characteristic frequency obtained in step 3 and rail resonance-characteristic frequency are established with step 5
Wheel rail relation database be compared；It is specific as follows：
601, when the difference of Wheel Rail Vibration characteristic frequency and the Wheel Rail Vibration characteristic frequency in wheel rail relation database is pre-
If in allowable range, and rail resonance-characteristic frequency and the gap of the rail resonance-characteristic frequency in wheel rail relation database also exist
In default allowable range；It is then that observation state is stored in wheel rail relation database by this group of data markers；
602, when the difference of the Wheel Rail Vibration characteristic frequency in Wheel Rail Vibration characteristic frequency and wheel rail relation database does not exist
The gap of rail resonance-characteristic frequency in default allowable range or in rail resonance-characteristic frequency and wheel rail relation database
It is that train wheels go wrong if the Wheel Rail Vibration characteristic frequency of whole railway changes when not in default allowable range,
Otherwise, then it is that rail goes wrong, is that alarm condition is stored in database, and train wheels are gone wrong by this group of data markers
Information, the information that goes wrong of rail be transferred to monitoring personnel；
Step 7: the characteristic frequency of step 4 ambient noise and data in background noise data library are compared and analyzed,
Big data analysis is carried out to background noise data, situation of change of the rail with external environment is analyzed, to obtain its aging rule
With bimetry information.
Fig. 2 is the spatial relation graph of optical fiber and rail of the present invention, and distributed optical fiber vibration sensing system centering is used in Fig. 2
The telecommunication optical fiber that state's line of high-speed railway is laid with is detected, and obtains a large amount of fiber-optic vibration curve data, data are shown as
Gray scale indicates the Waterfall plot of intensity, and abscissa is fiber lengths information and ordinate is time span information between curve.It carries
Take the track wherein generated by train driving, and according to before track, interested area division behind in track and track.Fig. 3 is Φ-
OTDR measures original signal figure, is the trajectory diagram of two train drivings as shown in Fig. 3, wherein abscissa is distance, indulges and sits
Mark is the time, and the track that a train driving is crossed can be indistinctly seen in figure, wherein including rail in two regions up and down
Vibration information.Region is divided according to the distance between every group of wheel in the track of train driving in Fig. 3, can obtain such as Fig. 4
Schematic diagram, Fig. 4 is processing rank rear wheel paths and front and back administrative division map, the health letter of the wheel comprising corresponding compartment in each region
Breath.
Train driving track is extracted, using the Boundary extracting algorithm based on Dynamic Programming.With based on second order zero crossing
Edge detection, compared with the edge extracting method based on search, have the characteristics that insensitive to noise in image, that is, carried out
When edge extracting, it is typically necessary and noise reduction filtering is carried out to reduce the erroneous judgement in edge extracting to artwork in advance.
Dynamic Programming (Dynamic Programming-DP) is a kind of effective method for solving shortest path in graph theory,
This method is put forward for the first time in early 1950s by American R.E.Bellman, and core views are can be the multistage
Decision process is solved one by one after being converted into multiple single phase problems, is that can obtain multistage decision using this method
The optimal solution of journey.DP algorithm can also be applied in the edge extracting problem in image procossing, and key is that this is asked
Topic is converted to a critical path problem, is then solved by backtracking method with DP algorithm.Marginal position is positioned using DP algorithm
Method can be divided into two steps：It establishes figure and solves shortest path：
1) figure is established：Using each pixel on image I as node, to arrange as layer, in each pixel and next layer with
Its line closed between three nearest pixels (image up-and-down boundary is two) is side.A weights are distributed for each side,
Respectively：
cost1(i, j)=λ w (I (i-1, j+1), I (i, j))+Cost (i-1, j+1),
cost2(i, j)=w (I (i-1, j), I (i, j))+Cost (i-1, j),
cost3(i, j)=λ w (I (i-1, j-1), I (i, j))+Cost (i-1, j-1),
Wherein λ is scale factor, and i, j respectively represent the directions x and the directions y of image, and
W (a, b)=2 × max (I)-a-b
After comparing these three weights, minimum weights are assigned to present node so that there are one MINIMUM WEIGHTs for present node tool
Value：
What the weights represented is the minimum weights (shortest distance) that the node is transmitted to by preceding layers.Calculating this most
While small weights, it is also necessary to for another parameter of each nodes records：
The reference record obtains all possible path of shortest path.Therefore, in the figure finally obtained, have at node
Two parameters, weights Cost and path P ath.
2) shortest path is solved：By building the process of figure it is found that last row has recorded the left side to right side of this figure in figure
The length of all possible shortest path, therefore the unique shortest path of solution needs position minimum total weight value in last row first
Then Cost recalls adjacent node one by one from back to front again.Trace-back process is determined by the parameter Path of above-noted, that is, is worked as
What the Path of front nodal point was recorded is the node coordinate that last layer should recall.Since solution procedure is successively to recall formula,
It ensure that the continuity at edge, can effectively exclude interference of the noise to lines.
Fig. 5 a are wheel track frequency relation figures, and Fig. 5 b are rail characteristic frequency figures.Data such as Fig. 5 a of certain position are chosen, are carried out
Data prediction vibrates two different vibrations such as the mode of Short Time Fourier Transform for rail independent resonant Wheel Rail Contact
It is handled.
Short Time Fourier Transform is a kind of research highly effective method of non-stationary signal, it is established becomes in conventional Fourier
On the basis of changing, basic thought is to introduce a window function γ (t) with time-frequency locality, it is allowed to be slided along signal, right
The each section of signal by window interception carries out Fourier transformation, can since the position of sliding window function introduces the information of time
To obtain the frequency analysis of a time-varying as a result, the Short Time Fourier Transform of signal s (t) in this way is defined as
* represents the conjugation of plural number in formula, and ω is pulse angular frequency, short time-window γ*(τ-t) is effectively limited in signal point
Analyse the output nearby of time τ=t.Short Time Fourier Transform is the frequency spectrum of a parts of the signal s (t) near time t.For two
Kind special circumstances, when window function selects Dirac function：
Short Time Fourier Transform has extraordinary temporal resolution at this time, but is difficult to provide preferable frequency discrimination
Rate, when window function selects normal function, Short Time Fourier Transform becomes the Fourier transformation of signal at this time, it has very
Good frequency resolution, but it does not provide any time resolution ratio:
Short Time Fourier Transform has many advantageous properties, it have linear TIME SHIFT INVARIANCE, frequency displacement invariance, band the general character,
The distribution situation for square representing energy of the signal on time frequency plane for the mould that low pass, calculation amount are small, its value takes, these properties
Theoretical foundation is provided for research light pulse signal transmission.
Fig. 5 b can be obtained by Short Time Fourier Transform, abscissa is the time in figure, and ordinate is frequency.In addition to can be with
It is clearly apparent outside the track of two trains, it is further seen that some apparent rail resonant frequency points.
Relatively accurate resonant frequency in order to obtain, using the fitting of a polynomial based on least square method (LM algorithms).LM
Algorithm is a kind of iterative algorithm solving non-linear real number function of many variables local minimum, can be regarded as steepest descent method and height
This-combination of Newton method, the local convergence of existing Gauss-Newton method, and have the global property of steepest descent method.LM algorithms
It is required that given wait for matched curve coefficient a1, a2, a3Initial guess value, it is initial to guess in the case where maximum iteration is constant
Value should be as possible close to the required initial value of best fit parameters.
In being estimated model, when such as to be estimated parameter and function be nonlinear relationship, will be transformed into one it is nonlinear
Least Square Solution.For the Parameter Estimation Problem of non-linear known relation formula, two methods are generally used at present, one is
Gauss-newton method, there are one be exactly LM algorithms.The parameter in nonlinear model to be estimated is determined using gauss-newton method, no
The fitting precision that the optimization of parameter and the parameter for avoiding adjusting parameter value repeatedly, and obtaining only may be implemented is also higher.But
Gauss-newton method there are disadvantage, be exactly given parameter initial value if inappropriate, generate after iteration function acts on and do not receive
The case where holding back characteristic, just will appear diverging.And LM algorithms are steepest descent method and the product that gauss-newton method is combined, and are in height
The algorithm that damping factor develops, therefore its existing local convergence characteristic are introduced on the basis of this Newton method, and also there is the overall situation
Characteristic.Its main thought is to solve for the local minimum of the non-linear real number function of many variables.Parameter value is solved using LM algorithms
When, it is desirable that an initial value is first set to the parameter in model, under the conditions of maximum iteration is immovable, it is desirable that initial value
It should be as close as the optimum value of parameter.
The general form of expression of nonlinear relation is：
Y=f (x1,x2...xi；a1,a2...ai)+ε
In formula, f is known nonlinear function, x1,x2…xiIndicate i independent variable, a1,a2...aiIn representative function
There are n unknown parameters to be estimated, ε to indicate stochastic error.The main thought of LM algorithms is to obtain the linearisation near certain point
Iterative formula gradually acquires optimal solution to carry out a series of interative computation.MakeLM algorithms have
Body realizes that steps are as follows：
A. it is assumed that akInitial value beThe initial approximate error Q of observation0For：
B. basis：
Calculate bij,biyThe initial value of given d simultaneously.
C. equation group is solved
And by akIt is revised as：
D. the approximate error Q between f and y is calculated againi：
E. a will be changedkFront and back error Qi-1With error QiSize comparison is carried out, if Qi-1< QiIt obtainsIf Qi-1
> QiIt obtains and needs to improve d value sizes, repeat step c, step d and step e.
F. step b, step c, step d and step e are constantly repeated always, | Δk| until value allows to miss less than specified
Difference just completes LM algorithms.
From the realization step of LM algorithms as can be seen that LM algorithms to not only avoid repeatedly value of adjusting parameter etc. a series of
Tedious work, and because the introducing of damping factor d relaxes the limitation to initial value, obtained numerical value is best fit parameters,
Fitting precision is set to improve many.
The resonance dot frequency that fitting is obtained vibrates two kinds of vibration modes according to for rail independent resonant and Wheel Rail Contact
And background noise classification preserves when without vehicle, classification carries out data and builds library, as the firsthand information of health comparison, typically
Rail resonance-characteristic frequency is illustrated in figure 6 along rail distribution map.
In monitoring process, Rayleigh rayleigh backscattering optical signal is carried out by the way of the above Short Time Fourier Transform
Analysis, obtains frequency diagram, is analyzed useful signal and extracted, extract what rail independent resonant and wheel were contacted with rail
Health data comparative analysis in two kinds of vibration frequencies, with database, respectively to both the above vibrate carry out discriminatory analysis, find with
There is the position of inconsistent vibration frequency in database, to realize to rail, roadbed and the real-time health monitoring of wheel, such as Fig. 7
Shown, the faint change of rail characteristic frequency occurs for same position, and basic frequency is constant, but secondary frequencies die down, and it is small to illustrate that rail occurs
Problem causes strain to change, does not have much affect to rail through railway department's investigation for temperature change.
The above content is a further detailed description of the present invention in conjunction with specific preferred embodiments, and it cannot be said that
The specific implementation of the present invention is confined to these explanations.For those of ordinary skill in the art to which the present invention belongs, exist
Under the premise of not departing from present inventive concept, several simple deductions or replacement can also be made, all shall be regarded as belonging to the present invention's
Protection domain.
Claims (7)
1. a kind of railway health monitor method based on distributed optical fiber vibration sensing system, which is characterized in that including following step
Suddenly：
Step 1: being sensed using the existing telecommunication optical fiber that Along Railway is laid with, capture Along Railway is respectively in train process
Two dimensional Distribution situation of the vibration information afterwards along time and space；By measured obtained fiber-optic vibration curve data countershaft on time
It is organized as the Waterfall plot that a gray scale indicates intensity, the abscissa of the Waterfall plot is fiber lengths information, ordinate is time span
Information；
Step 2: to there is the Waterfall plot corresponding to the period that train crosses to be filtered, using the edge based on Dynamic Programming
Extraction algorithm extraction train crosses the trajectory diagram of tested rail, includes Wheel Rail Vibration relation data in trajectory diagram；In trajectory diagram
Former and later two ranges of track are marked off, include rail vibration data within the scope of the two；
Step 3: carrying out time frequency analysis processing respectively to Wheel Rail Vibration relation data and rail vibration data, extracts wheel track and shake
Wheel Rail Vibration characteristic frequency spectrum and rail resonance-characteristic frequency spectrum are passed through spy by dynamic characteristic frequency spectrum and rail resonance-characteristic frequency spectrum respectively again
Frequency spectrum extraction algorithm is levied, Wheel Rail Vibration characteristic frequency and rail resonance-characteristic frequency are obtained；
Step 4: the Waterfall plot corresponding to the period crossed to no train carries out time-domain analysis processing, ambient noise is extracted
Characteristic frequency spectrum obtains the characteristic frequency of ambient noise by the characteristic frequency spectrum of ambient noise using characteristic frequency spectrum extraction algorithm；
Step 5: it is undamaged on the surface of rail and wheel, step 1 is repeated to step 4, establishes wheel rail relation
Database and background noise data library；By Wheel Rail Vibration characteristic frequency and rail resonance-characteristic frequency and corresponding train model and
Rail road section information is stored in wheel rail relation database；The characteristic frequency of ambient noise and rail road section information are stored in the back of the body
Scape noise database；
Step 6: the wheel that the Wheel Rail Vibration characteristic frequency obtained in step 3 and rail resonance-characteristic frequency are established with step 5
Rail relational database is compared；It is specific as follows：
601, when the difference of the Wheel Rail Vibration characteristic frequency in Wheel Rail Vibration characteristic frequency and wheel rail relation database permits default
Perhaps in range, and rail resonance-characteristic frequency and the gap of the rail resonance-characteristic frequency in wheel rail relation database are also being preset
In allowable range；It is then observation state by the Wheel Rail Vibration characteristic frequency obtained in step 3 and rail resonance-characteristic frequency marker
It is stored in wheel rail relation database；
602, when the difference of the Wheel Rail Vibration characteristic frequency in Wheel Rail Vibration characteristic frequency and wheel rail relation database is not default
The gap of rail resonance-characteristic frequency in allowable range or in rail resonance-characteristic frequency and wheel rail relation database does not exist
It is that train wheels go wrong if the Wheel Rail Vibration characteristic frequency of whole railway changes when in default allowable range, it is no
Then, then it is that rail goes wrong, is by the Wheel Rail Vibration characteristic frequency obtained in step 3 and rail resonance-characteristic frequency marker
Alarm condition is stored in wheel rail relation database, and the information that information that train wheels go wrong, rail are gone wrong is transmitted
To monitoring personnel；
Step 7: carrying out data in the characteristic frequency of the ambient noise obtained in step 4 and background noise data library to score
Analysis carries out big data analysis to background noise data, and analysis rail is advised with the situation of change of external environment to obtain its aging
Rule and bimetry information.
2. a kind of railway health monitor method based on distributed optical fiber vibration sensing system according to claim 1, special
Sign is that the filtering method in the step 2 is wavelet decomposition denoising.
3. a kind of railway health monitor method based on distributed optical fiber vibration sensing system according to claim 1, special
Sign is that the method for time frequency analysis processing is Short Time Fourier Transform or wavelet analysis.
4. a kind of railway health monitor method based on distributed optical fiber vibration sensing system according to claim 1, special
Sign is that characteristic frequency spectrum extraction algorithm is the polynomial fitting method based on least square method.
5. a kind of railway health monitor method based on distributed optical fiber vibration sensing system according to claim 1, special
Sign is, is sensed to the existing telecommunication optical fiber that Along Railway is laid with using distributed optical fiber vibration sensing system.
6. a kind of railway health monitor method based on distributed optical fiber vibration sensing system according to claim 1, special
Sign is that distributed optical fiber vibration sensing system is Φ-OTDR systems.
7. a kind of railway health monitor method based on distributed optical fiber vibration sensing system according to claim 1, special
Sign is that the track obtained in analytical procedure two realizes the positioning to train, measure speed in the position of spatial axes and time shaft
Degree and acceleration.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610688251.2A CN106323442B (en) | 2016-08-18 | 2016-08-18 | A kind of railway health monitor method based on distributed optical fiber vibration sensing system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610688251.2A CN106323442B (en) | 2016-08-18 | 2016-08-18 | A kind of railway health monitor method based on distributed optical fiber vibration sensing system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN106323442A CN106323442A (en) | 2017-01-11 |
| CN106323442B true CN106323442B (en) | 2018-11-13 |
Family
ID=57744106
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610688251.2A Active CN106323442B (en) | 2016-08-18 | 2016-08-18 | A kind of railway health monitor method based on distributed optical fiber vibration sensing system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106323442B (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| RU2794238C1 (en) * | 2021-12-21 | 2023-04-13 | Акционерное общество "Рязанское производственно-техническое предприятие "Гранит" | Method for operational control of railway stock location, its speed and integrity |
Families Citing this family (20)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107727227B (en) * | 2017-09-30 | 2019-05-21 | 南京大学 | Ultra-high-tension power transmission line icing based on Φ-OTDR waves monitoring method |
| JP7192879B2 (en) * | 2018-12-03 | 2022-12-20 | 日本電気株式会社 | Track monitoring system, track monitoring device, track monitoring method, and program |
| CN112014593B (en) * | 2019-05-28 | 2022-05-17 | 浙江德盛铁路器材股份有限公司 | Device and method for monitoring and evaluating quality condition of railway track basic equipment |
| CN112016180B (en) * | 2019-05-28 | 2023-12-15 | 浙江德盛铁路器材股份有限公司 | Quality condition analysis method for railway track foundation equipment |
| CN110579272A (en) * | 2019-07-30 | 2019-12-17 | 安徽工和环境监测有限责任公司 | Method for monitoring inspection and acceptance noise of railway special line |
| CN110631681A (en) * | 2019-09-26 | 2019-12-31 | 华北水利水电大学 | High-speed rail safe operation detection method based on optical fiber distributed vibration monitoring |
| CN110686626B (en) * | 2019-10-14 | 2021-03-19 | 南京大学 | OPGW (optical fiber composite overhead ground wire) icing monitoring system and method based on phase-sensitive optical time domain reflection system |
| CN110779614B (en) * | 2019-10-31 | 2021-07-06 | 武汉理工光科股份有限公司 | Submarine cable anchor damage monitoring and positioning method and system based on distributed optical fiber sensing |
| CN111071300B (en) * | 2020-02-12 | 2021-12-14 | 太原理工大学 | High-speed train rail transit fault safety monitoring and early warning system and signal processing method |
| CN111766573B (en) * | 2020-06-02 | 2024-02-27 | 武汉烽理光电技术有限公司 | Method and system for improving array grating positioning spatial resolution through Kalman filtering |
| CN111845843A (en) * | 2020-07-13 | 2020-10-30 | 南京大学 | High-speed rail safe operation detection method based on optical fiber distributed vibration monitoring |
| CN111854919B (en) * | 2020-07-16 | 2021-09-17 | 南京大学 | Railway health monitoring method based on distributed optical fiber vibration sensing system |
| CN112033674B (en) * | 2020-08-07 | 2022-03-15 | 太原理工大学 | Train fault online detection system and method |
| CN112550378B (en) * | 2020-11-26 | 2023-03-17 | 安徽徽一通讯科技有限公司 | Method, system and storage for mapping positions between track accompanying optical cable and track |
| CN112611442A (en) * | 2020-11-26 | 2021-04-06 | 中国科学院合肥物质科学研究院 | Railway bridge health monitoring method and system based on track accompanying optical cable |
| CN112433004A (en) * | 2020-12-02 | 2021-03-02 | 长安大学 | Wheel rail damage monitoring method and system |
| CN114194255B (en) * | 2021-12-31 | 2022-07-19 | 安徽洪鸿光电技术有限公司 | Train locking tracking method based on optical fiber sensing technology |
| CN114518163B (en) * | 2022-02-21 | 2024-03-19 | 无边界(苏州)新材料科技有限公司 | Method for full acoustic optical fiber monitoring based on Gaussian-LM algorithm |
| CN117002556A (en) * | 2022-04-28 | 2023-11-07 | 中兴通讯股份有限公司 | Detection method and device applied to railway track and detection equipment |
| CN115384581A (en) * | 2022-10-27 | 2022-11-25 | 高勘(广州)技术有限公司 | Track monitoring method, device, equipment and storage medium |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN201980252U (en) * | 2010-12-16 | 2011-09-21 | 西安金和光学科技有限公司 | Rail monitoring and sensing device |
| CN102530028A (en) * | 2012-01-19 | 2012-07-04 | 重庆安谐新能源技术有限公司 | Railway track real-time monitoring system and data processing method thereof |
| CN202491817U (en) * | 2012-01-19 | 2012-10-17 | 重庆安谐新能源技术有限公司 | Real-time monitoring system for railway track |
| CN103215865A (en) * | 2013-04-03 | 2013-07-24 | 武汉理工大学 | Method for dynamically monitoring health states of railway continuous-beam bridge and rail overlapping device |
| CN104374465A (en) * | 2014-11-18 | 2015-02-25 | 浙江大学城市学院 | Subway low-frequency vibration test device and method based on fiber bragg grating sensors |
| KR20150111888A (en) * | 2015-09-08 | 2015-10-06 | 한국철도기술연구원 | A railway vehicles turck equipped with the direct drive in carbon fiber fabric coupling |
| CN105608822A (en) * | 2015-12-25 | 2016-05-25 | 中国铁道科学研究院电子计算技术研究所 | Railway perimeter intrusion vibration fiber alarm method and device |
-
2016
- 2016-08-18 CN CN201610688251.2A patent/CN106323442B/en active Active
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN201980252U (en) * | 2010-12-16 | 2011-09-21 | 西安金和光学科技有限公司 | Rail monitoring and sensing device |
| CN102530028A (en) * | 2012-01-19 | 2012-07-04 | 重庆安谐新能源技术有限公司 | Railway track real-time monitoring system and data processing method thereof |
| CN202491817U (en) * | 2012-01-19 | 2012-10-17 | 重庆安谐新能源技术有限公司 | Real-time monitoring system for railway track |
| CN103215865A (en) * | 2013-04-03 | 2013-07-24 | 武汉理工大学 | Method for dynamically monitoring health states of railway continuous-beam bridge and rail overlapping device |
| CN104374465A (en) * | 2014-11-18 | 2015-02-25 | 浙江大学城市学院 | Subway low-frequency vibration test device and method based on fiber bragg grating sensors |
| KR20150111888A (en) * | 2015-09-08 | 2015-10-06 | 한국철도기술연구원 | A railway vehicles turck equipped with the direct drive in carbon fiber fabric coupling |
| CN105608822A (en) * | 2015-12-25 | 2016-05-25 | 中国铁道科学研究院电子计算技术研究所 | Railway perimeter intrusion vibration fiber alarm method and device |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| RU2794238C1 (en) * | 2021-12-21 | 2023-04-13 | Акционерное общество "Рязанское производственно-техническое предприятие "Гранит" | Method for operational control of railway stock location, its speed and integrity |
Also Published As
| Publication number | Publication date |
|---|---|
| CN106323442A (en) | 2017-01-11 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106323442B (en) | A kind of railway health monitor method based on distributed optical fiber vibration sensing system | |
| US11562224B2 (en) | 1D-CNN-based distributed optical fiber sensing signal feature learning and classification method | |
| CN105139086B (en) | Track transition Amplitude Estimation method based on optimization confidence rule-based reasoning | |
| US20180080812A1 (en) | Distributed optical fiber sensing signal processing method for safety monitoring of underground pipe network | |
| CN101556724B (en) | Safety management system of optical fiber perimeter and pattern recognition method thereof | |
| Xue et al. | Pothole in the dark: Perceiving pothole profiles with participatory urban vehicles | |
| CN104729667B (en) | A kind of disturbance kind identification method in distributed optical fiber vibration sensing system | |
| CN104240455A (en) | Method for identifying disturbance event in distributed type optical fiber pipeline security early-warning system | |
| Bai et al. | Detection and identification of external intrusion signals from 33 km optical fiber sensing system based on deep learning | |
| CN107588915A (en) | A kind of Bridge Influence Line recognition methods and system | |
| CN104217513B (en) | The method improving phase sensitive optical time domain reflectometer identification intrusion event accuracy rate | |
| CN111222743B (en) | Method for judging vertical offset distance and threat level of optical fiber sensing event | |
| CN111497902A (en) | phi-OTDR (optical time Domain reflectometer) -based high-speed rail health online monitoring method | |
| Yang et al. | Pipeline safety early warning by multifeature-fusion CNN and LightGBM analysis of signals from distributed optical fiber sensors | |
| Fouda et al. | Pattern recognition of optical fiber vibration signal of the submarine cable for its safety | |
| Aboah et al. | Smartphone-based pavement roughness estimation using deep learning with entity embedding | |
| DE112021004482T5 (en) | DETECTION AND LOCALIZATION OF ACOUSTIC IMPULSE AT CITY SCALE | |
| CN110455397A (en) | A kind of shop equipment failure optical fiber sensing method and system based on LSTM | |
| CN112550369A (en) | Switch application on-line state monitoring system | |
| Cui et al. | Identifying accident black spots based on the accident spacing distribution | |
| CN105629257A (en) | Power transmission line distributed vertical span monitoring device and method | |
| Nedoma et al. | Fiber-optic interferometric sensor for monitoring automobile and rail traffic | |
| US20230024104A1 (en) | Identification of false transformer humming using machine learning | |
| CN106646670A (en) | Distributed micrometeorological monitoring method for transmission lines | |
| Cai et al. | Train detection and classification using distributed fiber-optic acoustic sensing |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| CP03 | Change of name, title or address | ||
| CP03 | Change of name, title or address |
Address after: 210000 No. 2 Xianyao Road, Yaohua Street, Qixia District, Nanjing, Jiangsu Province Patentee after: Nanjing Faab Photoelectric Technology Co., Ltd. Address before: 210000 Room 1002, 16 Building, Landscape Beiyuan, Shangcheng, No. 7 Yaojia Road, Yaohua Street, Qixia District, Nanjing City, Jiangsu Province Patentee before: Nanjing Faaibo Optoelectronics Technology Co., Ltd. |