CN107063339A - Falling rocks judges recognition methods with train signal classification classification along railway - Google Patents
Falling rocks judges recognition methods with train signal classification classification along railway Download PDFInfo
- Publication number
- CN107063339A CN107063339A CN201710065942.1A CN201710065942A CN107063339A CN 107063339 A CN107063339 A CN 107063339A CN 201710065942 A CN201710065942 A CN 201710065942A CN 107063339 A CN107063339 A CN 107063339A
- Authority
- CN
- China
- Prior art keywords
- signal
- energy
- amplitude
- falling rocks
- time
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01H—MEASUREMENT OF MECHANICAL VIBRATIONS OR ULTRASONIC, SONIC OR INFRASONIC WAVES
- G01H17/00—Measuring mechanical vibrations or ultrasonic, sonic or infrasonic waves, not provided for in the preceding groups
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01D—MEASURING NOT SPECIALLY ADAPTED FOR A SPECIFIC VARIABLE; ARRANGEMENTS FOR MEASURING TWO OR MORE VARIABLES NOT COVERED IN A SINGLE OTHER SUBCLASS; TARIFF METERING APPARATUS; MEASURING OR TESTING NOT OTHERWISE PROVIDED FOR
- G01D21/00—Measuring or testing not otherwise provided for
- G01D21/02—Measuring two or more variables by means not covered by a single other subclass
Abstract
For prior art to the information extraction in vibration signal along railway using limited, the technological means of its multiple thresholds defect limited to improving discriminant accuracy effect judges recognition methods the invention provides falling rocks along a kind of railway and train signal classification.This method arithmetic center by I grades of signal after primary signal according to denoising, amplitude index sentence knowledge, II grades of the signal based on FFT sentence knowledge, knowledge is sentenced based on STFT III grade of the signals converted the step of implement, it is final to judge whether falling rocks occurs with driving a vehicle.The invention also discloses falling rocks energy scale computational methods, the current state monitoring method of driving.The present invention analyzes amplitude, energy spectrometer, frequency analysis integrally combine, and uses no signal analysis index and analysis method in knowledge rank aspect by sentencing in difference, realizes that the knowledge degree of accuracy is sentenced in raising.The principle of the invention is reliable, and calculating process science is easy, implements instrument simply, and easy for installation, results of measuring precision is high, the need for being particularly suitable for use in engineering field.
Description
Technical field
Judge that recognition methods, especially one kind are related to falling rocks along railway the present invention relates to a kind of classification classification of vibration signal
Sentence knowledge method with the classification of train passage signal, belong to vibration signal monitoring field of measuring technique, traffic signalization neck
Domain.
Background technology
Vibration signal refers to as the signal produced by non-static object.Because nonstatic state is the absolute status of object,
Thus vibration characteristics and the vibration signal of generation turn into the inherent characteristic of object.Under excited state, if the excitation of vibration source
When identical or close with the inherent characteristic parameter of object, the resonance response caused by the superposition of each frequecy characteristic information can be produced.
The temporal signatures of vibration signal are mainly reflected in the characteristics such as amplitude, cycle, phase, and its frequency domain character is then mainly manifested in frequency
In rate, energy information.Because different objects have the vibration signal of its characteristic, same object is also shown under different conditions
The vibration signal of different characteristic, therefore, by extracting and analyzing various spies that are original in vibration signal or being produced via conversion
Reference ceases, and can reversely calculate and position the object vibrated in itself, and its motion state.
Application publication number discloses a kind of pre- police of railway falling rocks for the A of CN 102079319 Chinese invention patent application
Method, its technical scheme comprises the following steps：The fiber grating that strain is produced with rail vibration is provided, fiber Bragg grating (FBG) demodulator is utilized
Device obtains the electric signal of the wavelength change for the light that reflection is reflected back from fiber grating, is obtained based on electric signal from fiber grating reflection
The wavelength variation information of the light returned, and determine whether based on the wavelength variation information to send alarm.The purpose is to utilize optical fiber light
Gate sensor has that electromagnetism interference, size be small, lightweight advantage.Although this method is provided with first threshold and Second Threshold
To detect that railway falling rocks occurs, judge that railway falling rocks occurs according to the magnitude relationship of amplitude and threshold value, amplitude variations speed
Whether.But have a disadvantage in that the electrical signal amplitude of the wavelength change for the light that multiple thresholds discriminant criterion only fiber grating is reflected back
Sole indicator, thus limited to the vibration information extraction and application in vibration signal, its use the technological means of multiple thresholds for
Improve and differentiate that the benifit of result accuracy is similarly limited.
The content of the invention
The purpose of the present invention is classified aiming at the deficiencies in the prior art there is provided falling rocks along a kind of railway and train signal
Recognition methods is judged, to realize the skill for being sentenced knowledge to the falling rocks harm along railway, traffic safety using analysis of vibration signal
Art purpose.Its technical scheme is as follows：
Falling rocks judges recognition methods with train signal classification along a kind of railway, and site occurs for potential falling rocks along railway
Arrangement vibrating sensor gathers, transmits primary signal to arithmetic center in real time, and arithmetic center obtains original time-amplitude signal simultaneously
Implement according to following steps：
Step S1, denoising
Flip-flop, long period skew, the low-frequency anomaly in primary signal are filtered out, filtered signal is obtained, into step
S2；
Step S2, I grades of vibration signal sentence knowledge
Amplitude threshold A is sets, as filtered signal amplitude >=AsWhen, it is I grades of vibration signals to sentence knowledge, records the moment value
T1, into step S3；The amplitude threshold AsDrawn by falling rocks vibration test result and/or historical summary data statistics result
Falling rocks vibration amplitude variation characteristic determine；
It is characterized in that：
Step S3, II grades of vibration signal sentence knowledge
Step S31, FFT
Filtered signal is subjected to FFT, obtained from moment value T1The frequency-amplitude signal risen, into step S32；
Step S32, signal sentence knowledge
Frequency-energy signal is obtained by frequency-amplitude signal of change obtained by step S31, energy ratio is calculated；The energy
Ratio is the ratio of signal energy value and whole signal energy values in 0Hz~500H frequency ranges in frequency-energy signal；
If energy ratio >=energy ratio threshold value RE, it is II grades of falling rocks vibration signals to sentence knowledge, records quarter time value T21, exit
Step S32 enters step S4；If energy ratio ＜ energy ratio threshold values RE, it is II grades of driving vibration signals to sentence knowledge, records the time
Quarter value T22, exit step S32 enter step S4；
The energy ratio threshold value RE=80%~90%；
Step S4, III grade of vibration signal sentence knowledge
Step S41, STFT is converted
Filtered signal is subjected to STFT conversion, obtained from moment value T21Or T22T/F-the energy signal risen, enters
Enter step S42；
Step S42, signal sentence knowledge
From quarter time value T21Or T22Rise along time shaft access time-Frequency point to the right, and calculate at the time of chosen-frequency
The energy value average of rate point, when energy value average >=average energy value threshold value A0When be determined as vibration, continue along time axial direction
Right access time-Frequency point, as the energy value average ＜ A in continuous 2s0When, it is judged as that vibration terminates, energy value average threshold value A0
=(1.5As)2；
Least square fitting is carried out to the energy value of the T/F point in the vibration beginning and ending time, this vibration phase is obtained
Between fitting coefficient set；If fitting coefficient set meets energy fit threshold LSF1When, it is III grade of falling rocks vibration signal to sentence knowledge,
Determine that moment T occurs for falling rocks31It is energy value average >=average energy value threshold value A0At the time of；If fitting coefficient set meets energy plan
Close threshold value LSF2When, it is III grade of driving vibration signal to sentence knowledge, it is determined that driving is by occurring moment T32It is energy value average ＜ energy
Average threshold value A0At the time of；
The energy fit threshold LSF1Be in fitting coefficient set 60% and the above fitting coefficient≤- 3, the energy
Fit threshold LSF2Be in fitting coefficient set 60% and the above fitting coefficient ＞ -3；
Judge whether falling rocks occurs according to III grade of falling rocks vibration signal identifying result, knowledge is sentenced according to III grade of driving vibration signal
As a result driving is judged whether.
Falling rocks is classified with train signal along above-mentioned railway judges that the basic fundamental design of recognition methods includes two aspects：One
Aspect filters out falling rocks vibration signal and driving two kinds of monitoring objects of vibration signal respectively from vibration signal, and sentencing knowledge, whether it sends out
It is raw；On the other hand the filtered rear progress of vibration signal gathered to sensor sentences knowledge according to amplitude, according to the classification of frequency-energy etc.,
Analysis identifying result precision is improved step by step.
The purpose for knowing precision is sentenced to realize to improve by Gradual Differentiating, when technical solution of the present invention is used in vibration signal
The frequency in amplitude information and frequency domain, energy information in domain carry out the mode of substep computing.Classification is sentenced in knowledge, to filtering and noise reduction
Vibration signal judges identification using three-level after filtering afterwards：(1) it is that vibration amplitude sentences knowledge that I grades, which are sentenced knowledge,.Amplitude is that vibration signal exists
Most basic Faults by Vibrating in time domain, knowledge is sentenced using vibration, can tentatively be sentenced and be known under the state whether have high-magnitude vibrations
Event causes the mutation of signal, excludes basic ambient noise.By setting the lower threshold of vibration amplitude, tentatively sentence knowledge this when
Whether quarter has train to pass through or falling rocks event.But amplitude sentences that knowledge method is excessively rough, it is only capable of qualitative sentencing the big of knowledge event
General possibility, thus knowledge vibration signal sentenced using amplitude exist ambient noise, instrument error itself are mistaken for validity event
Shortcoming, to overcome this defect, it is necessary to which signal transacting further quantitatively sentences knowledge.(2) it is the frequency domain frequency based on FFT that II grades, which are sentenced knowledge,
Rate-energy sentences knowledge.The typical use of FFT is to compose signal decomposition into frequency-amplitude, and so we are it is known that vibrations
Each frequency content and its amplitude of signal, and then obtain Energy distribution situation of the signal in each frequency range.Pass through FFT
Quantitative analysis, the Energy distribution situation for each frequency range for obtaining event signal can be quantified, then with train signal, falling rocks signal
Frequency-energy feature screen, sentence and know falling rocks event or train event.Sentencing knowledge for II grades can be by FFT
The frequency distribution of vibration signal is obtained, can quantify and obtain the frequency range that event signal main component is concentrated, but this method lacks
The temporal information of signal, the only frequency of signal-energy information defect are lost, it is therefore desirable to which further III grade is sentenced knowledge.(3)Ⅲ
It is to sentence knowledge based on the STFT frequency-energy converted that level, which is sentenced and known,.STFT conversion is a kind of two-dimentional time-frequency conversion, can obtain vibrations letter
Number time-frequency domain Energy distribution, the vibration signal of different event has significant difference in time-frequency domain, can pass through
Vibration signal is quantitatively calculated to carry out sentencing knowledge in the distribution and intensity of time-frequency domain energy.Sentenced by III grade and know STFT's
Quantitative analysis, can quantify the Energy distribution situation for each frequency range for obtaining vibration event signal, then believed with train signal, falling rocks
Number frequency-energy feature screen, sentence the type (i.e. falling rocks event or event of driving a vehicle) and temporal characteristics for knowing outgoing event.This
Technical scheme sentences the basic skills known and all compared using threshold value in being recognized to the judgement step by step of vibration signal per one-level.I grades
Knowledge is sentenced according to amplitude threshold As, II grades are sentenced knowledge according to energy ratio threshold value RE, III grade is sentenced knowledge according to average energy value threshold value A0, energy
Fit threshold etc..These threshold values are drawn by the falling rocks of early stage or driving vibration test interpretation of result, or according to falling
What stone or driving vibration historical summary record statistical analysis were drawn.
Falling rocks vibration signal and the purpose of driving two kinds of monitoring objects of vibration signal are filtered out to realize, technical scheme is main
Pass through two class technological means：One is that threshold value is divided both amplitude class threshold values, energy class threshold value, frequency class threshold value when setting
Or a setting She Zhi not be selected according to condition；Two be that the analytic operation based on frequency partition is carried out to signal.
Usually, in step S1, using the FIR of the rank of 6 ranks~10 to original time-amplitude signal denoising.Amplitude threshold As
It is background signal amplitude Ab1.5 times, the background signal amplitude AbIt is initial under nature after vibrating sensor in-site installation
Signal amplitude average value within 1min.
Falling rocks judges that recognition methods is solved with train signal classification and sentences knowledge falling rocks according to vibration signal along above-mentioned railway
The problem of generation and train passage.Based on this, the present invention further provides a kind of falling rocks energy scale based on its realization
Computational methods.Its technical scheme is as follows：
It is a kind of to be classified the rock-fall impact energy for judging that recognition methods is realized with train signal using falling rocks along above-mentioned railway
Scale computational methods, it is characterised in that：
In the step S42, after sentencing knowledge for III grade of falling rocks vibration signal, quarter time value T is recorded31, exit step S42 enters
Enter step S5A；
Step S5A, falling rocks energy scale computational methods
Step S5A1, impact energy simulation calculating beginning and ending time sentence knowledge
Using STA/LTA methods to T1Filtered signal carries out Vibration identification from moment；As STA/LTA ratios >=STA/
LTA fractional thresholds RSLWhen be defined as falling rocks Energy Simulation zero computing time t1, as amplitude≤amplitude lower threshold AminWhen determine
End time t is calculated for falling rocks Energy Simulation2, exit step S5, into step S6；
The short cycle t of STA/LTAS, STA/LTA long periods tL, STA/LTA fractional thresholds RSLPass through falling rocks vibration test result
And/or the falling rocks vibration amplitude variation characteristic that historical summary data statistics result is drawn is determined, amplitude lower threshold AminBy falling
The falling rocks vibration amplitude variation characteristic that stone vibration test result and/or historical summary data statistics result are drawn is determined；
Step S5A2, calculating falling rocks scale
Falling rocks, which is calculated, according to formula 5.1 tumbles the scale A that releases energyE；
In formula, E (t) --- moment t1Square of~moment t RMS amplitude,
E0--- noise level, calculate and determine according to formula 5.2,
In formula, A1--- time t1Locate amplitude,
A2--- time t2Locate amplitude.
Above-mentioned rock-fall impact energy method computations are to be classified in falling rocks vibration signal in the result for judging identification, using STA/
LTA methods are further shaken facies analysis.Because STA/LTA algorithms can identify the terminal and its envelope energy of falling rocks signal,
Energy integral thus is carried out the temporal frequency point start/stop time, you can assess falling rocks scale, it is thus possible to more accurately true
Surely start-stop state change is vibrated.The STA/LTA fractional thresholds R being related in STA/LTA analysesSL, amplitude lower threshold AminEqually
The experiment that can be vibrated by early stage falling rocks and/or the result of historical data analysis draw determination.The STA/LTA that the present invention is provided
Preferred parameter in analysis method is：
In STA/LTA methods, short period time tS=0.05s, long period time tL=1s, STA/LTA fractional threshold RSL
=3~3.8, window function ω [n] are the Hanning windows that length is 128, are calculated according to formula 5.3,5.4.Amplitude lower threshold Amin=10%
Amax~15%Amax, AmaxIt is t1The maximum of amplitude in~moment t time interval.
Falling rocks judges that recognition methods is solved with train signal classification and sentences knowledge row according to vibration signal along above-mentioned railway
On the basis for the problem of car is current, knowledge side is sentenced the present invention further provides a kind of driving based on its realization current beginning and ending time
Method.Its technical scheme is as follows：
It is classified using falling rocks along above-mentioned railway and train signal when judging the current start-stop of driving that recognition methods is realized
Between sentence knowledge method, it is characterised in that：
In the step S42, after sentencing knowledge for III grade of driving vibration signal, quarter time value T is recorded32, exit step S42 enters
Enter step S5B, the current start-stop of driving and sentence knowledge；
The step S5B8 includes：As amplitude ＜ train travel amplitude thresholds ALWhen be defined as drive a vehicle end time T4；It is described
Train travel amplitude threshold ALIt is the vibration letter that vibrating sensor measurement is obtained when being travelled with pre-set velocity under train Light Condition
Number amplitude.
It is directly using real-time amplitude and train travel amplitude threshold A that the above-mentioned driving current beginning and ending time, which sentences knowledge method,LThan
To method determine train whether in driving pass through state.
Sentence knowledge method using the above-mentioned driving current beginning and ending time, whether can also further solve the current state of monitoring driving
Technical problem.Its technical scheme is as follows：
A kind of current state monitoring method of driving that the realization of knowledge method is sentenced using the above-mentioned driving current beginning and ending time, it is special
Levy and be：It is determined that driving end time T4Sentence knowledge into step S6, the current state of driving afterwards；The step S6 includes：
From moment value T32Rise along time shaft access time-Frequency point and to the amplitude of selected T/F point to the right
Carry out least square fitting, whenever fitting coefficient be less than 0 when count is incremented, and continue along time shaft move right fitting until
Moment value T4；Finally it is counted as N, compartment joint number=(N+1)/2；
Overall length of the train degree is obtained according to compartment joint number and single-unit railway car length computation, according to overall length of the train degree and driving
Transit time, which is calculated, obtains train travel speed, and judges whether driving passage rate is normal.
Above-mentioned driving is passed through in state monitoring method, is the part feelings according to track interior to the principle of compartment method of counting
Condition, we are analyzed using the situation of Train Track to carry out.Numerical-Mode is carried out to stock rail Space Coupling kinetic model
Intend additional actual monitoring, show that every process one saves train, it is first that the shape facility of a spindle, i.e. vibration energy can be presented in signal
Reduced after increasing, crest value projection, therefore we just can recognize that the section of train according to the number of spindle on signal spectrum
Number.
Compared with prior art, the beneficial effects of the invention are as follows：(1) the invention provides falling rocks and row along a kind of railway
The classification of car Modulation recognition judges recognition methods.This method analyzes amplitude, energy spectrometer, frequency analysis integrally combine, and passed through
Sentence in difference and use no signal analysis index and analysis method in knowledge rank aspect, realize the mesh for improving and sentencing and knowing the degree of accuracy
's.It is this that various analysis is used in combination and is intended to improve train falling rocks and train signal analysis precision along railway, especially
It is by techniqueflow different phase use different analysis methods, it is comprehensive to realize that knowledge is sentenced in the classification to two class vibration signals
The technical concept for closing raising analysis precision did not occurred in the prior art.(2) the invention provides real using preceding method
Existing falling rocks energy scale computational methods, can solve the problem that falling rocks along railway tumbles the measuring and calculating problem of energy scale, can be applied to
The Related Research Domain of geological disaster.(3) the invention provides the current state monitoring method of the driving realized using preceding method,
This method determines the train passage time using analysis of vibration signal, and counts railway carriage quantity, and then calculates train travel speed, can
Train traffic safety whether state monitored.
Brief description of the drawings
Fig. 1 sensor mounting location schematic diagrames.
Fig. 2 is arithmetic center algorithm flow schematic diagram.
Fig. 3 is falling rocks 1#, 2#, 3# event signal STFT spectrograms.
Fig. 4 is falling rocks 1# events primary signal, FFT spectrum and STFT spectrograms.
Fig. 5 is falling rocks 2# events primary signal, FFT spectrum and STFT spectrograms.
Fig. 6 is falling rocks 3# events primary signal, FFT spectrum and STFT spectrograms.
Fig. 7 is falling rocks 1# event STA/LTA arithmetic result figures.
Fig. 8 is falling rocks 2# event STA/LTA arithmetic result figures.
Fig. 9 is falling rocks 3# event STA/LTA arithmetic result figures.
Figure 10 is that train passes through primary signal, FFT spectrum and STFT spectrograms.
Embodiment
Below in conjunction with the accompanying drawings, the preferred embodiments of the present invention are further described.
Embodiment one
As shown in Fig. 1~Fig. 9, at certain Along Railway falling rocks easily hair monitoring falling rocks with the inventive method, a situation arises.
Site arrangement vibration acceleration sensor occurs for certain potential falling rocks along railway.Sensor is arranged on the interior of rail
On side or tie (five-pointed star shows sensing station in Fig. 1, figure).Sensor gathers, transmits primary signal to arithmetic center in real time,
Arithmetic center is obtained after original time-amplitude signal according to technical scheme flow implementation shown in Fig. 2.
Implementation below is demonstrated by taking the data processing of a sensor as an example.The Sensor monitoring to 3 falling rocks events,
Numbering is falling rocks 1#, 2#, 3# event (Fig. 3).Fig. 4 is falling rocks 1# events primary signal, FFT spectrum and STFT spectrograms, and Fig. 5 is
Falling rocks 2# events primary signal, FFT spectrum and STFT spectrograms, Fig. 6 are falling rocks 3# events primary signal, FFT spectrum and STFT
Spectrogram.
Step S1, denoising
Primary signal filters out flip-flop, long period skew, low-frequency anomaly in primary signal, obtains filtered signal.
Filtering can filter out low-frequency anomaly simultaneously, it is to avoid cause false-alarm and the interference to subsequent treatment.Filtering process is using 100 ranks
FIR is to original time-amplitude signal denoising.
Into step S2.
Step S2, I grades of vibration signal sentence knowledge
Measuring environment background signal amplitude Ab=0.13, amplitude threshold A is sets=1.5Ab=0.2.
By the amplitude of real-time judge sensor output signal, confirm that amplitude is more than in advance at 194.5s, 494.5s, 597s
If amplitude threshold As(Fig. 5), it is I grades of falling rocks vibration signals to sentence knowledge, records corresponding 3 moment value T1=194.5s, 494.5s,
597s, into step S3.
Step S3, II grades of vibration signal sentence knowledge
Step S31, FFT
Filtered signal is subjected to FFT, obtained from moment value T1The frequency-amplitude signal risen, into step S32；
Step S32, signal sentence knowledge
Frequency-energy signal is obtained by frequency-amplitude signal of change obtained by step S31, energy ratio threshold value R is setE=
80%.
Calculating is obtained at time 194.7s, 494.7s, 597.1s, and vibration frequency range 0Hz~500Hz energy=
250th, 260,200, whole signal energy value=300,305,220, energy ratio=83.3%, 85.2%, 91.0% are all higher than
RE, it is II grades of falling rocks vibration signals to sentence knowledge, and quarter time value T is recorded respectively21=194.7s, 494.7s, 597.1s, exit step
S32 enters step S4.
Step S4, III grade of vibration signal sentence knowledge
Step S41, STFT is converted
Filtered signal is subjected to STFT conversion, obtained from moment value T21T/F-the energy signal risen.STFT becomes
It is a kind of two-dimentional time-frequency conversion to change, and Energy distribution of the vibration signal in time-frequency domain is obtained by conversion.Due to falling rocks vibration
There is significant difference in time-frequency domain with other vibrations, thus can be by calculating the two distribution in time-frequency domain energy
Scope and intensity are recognized to falling rocks vibration signal.
Into step S42.
Step S42, signal sentence knowledge
Average energy value threshold value A is set0=(1.5As)2=0.09, energy fit threshold LSF is set1To be fitting coefficient collection
Close fitting coefficient≤- 3 of interior 60% and the above.
From quarter time value T21Rise along time shaft access time-Frequency point to the right, and calculate at the time of chosen-Frequency point
Energy value average, carries out signal and sentences knowledge.As a result such as Fig. 6 is shown, time 194.8s~195.3s, 494.8s~495.5s, 597s
~597.7s is the vibration beginning and ending time.Least square fitting is carried out to the energy value of the interval T/F point, and compares plan
Close coefficient sets and the relation of energy fit threshold.As a result show, 3 wayside signaling duration 60% and above fitting system
Number<- 3, it is satisfied by energy fit threshold LSF1, it is III grade of falling rocks vibration signal to sentence knowledge, and is confirmed in moment T31=194.9s,
495.0s, 597.4s there occurs that falling rocks is tumbled.
Into step S5A.
Step S5A1, Energy Simulation calculating beginning and ending time sentence knowledge
Using STA/LTA methods to T1Filtered signal carries out Vibration identification from moment；As STA/LTA ratios >=STA/
LTA fractional thresholds RSLWhen be defined as falling rocks Energy Simulation zero computing time t1, as amplitude≤amplitude lower threshold AminWhen determine
End time t is calculated for falling rocks Energy Simulation2；
In STA/LTA methods, short cycle t is setS=0.05s, long period tL=1s, STA/LTA fractional threshold RSL=3,
Window function ω [n] is the Hanning window that length is 128.Amplitude lower threshold A is setmin=10%Amax, AmaxThat is PGV in figure.
Result of calculation shows (Fig. 8), falling rocks Energy Simulation zero computing time t1=194s, 493s, 596.7s, correspondingly shake
Width A1=0.3,0.3,0.2, falling rocks Energy Simulation calculates end time t2=196.5s, 496s, 598.8s, corresponding amplitude A2=
0.2nd, 0.35,0.1, falling rocks tumbles cycle DUR=2.5s, 3s, 2.1s.
Fig. 7 is falling rocks 1# event STA/LTA arithmetic result figures, and Fig. 8 is falling rocks 2# event STA/LTA arithmetic result figures, Fig. 9
It is falling rocks 3# event STA/LTA arithmetic result figures.
Step S5A2, calculating falling rocks scale
Calculate to moment t1~moment t RMS amplitude E (t)=317.9,413.9,161.5, noise level E0=
0.0625th, 0.0977,0.0225, falling rocks tumbles the scale A that releases energyE=794.18J, 1240.7J, 338.9J.
The falling rocks of 1 embodiment of table one differentiates
Embodiment two
As shown in Figure 10, train passage situation is monitored at certain Along Railway falling rocks easily hair with the inventive method.This reality
Step S1~step S31 contents identical with embodiment one in mode are applied to be not repeated.From at step S32 at record data
Reason process.
Step S32, signal sentence knowledge
Frequency-energy signal is obtained by frequency-amplitude signal of change obtained by step S31, energy ratio threshold value R is setE=
80%.
Calculating is obtained at time 4.2s, vibration frequency range 0Hz~500Hz energy=2150, whole signal energies
Value=3000, the ＜ R of energy ratio=71.7%E, it is II grades of driving vibration signals to sentence knowledge, records quarter time value T22=4.2s, is moved back
Go out step S32 into step S4.
Step S4, III grade of vibration signal sentence knowledge
Step S41, STFT is converted
Filtered signal is subjected to STFT conversion, obtained from moment value T22T/F-the energy signal (Figure 10) risen.
Into step S42.
Step S42, signal sentence knowledge
Energy fit threshold LSF is set2Be in fitting coefficient set 60% and the above fitting coefficient >=-3.
From quarter time value T22Rise along time shaft access time-Frequency point to the right, and calculate at the time of chosen-Frequency point
Energy value average, carries out signal and sentences knowledge.As a result such as Fig. 8 is shown, time 4.4s~7.5s is the vibration beginning and ending time.To the interval
The energy value of T/F point carries out least square fitting, and compares the relation of fitting coefficient set and energy fit threshold.Knot
Fruit display, the fitting coefficient of wayside signaling duration 75%<- 3, meet energy fit threshold LSF2, it is III grade of driving to sentence knowledge
Vibration signal, and confirm train in moment T32=4.5s starts to pass through.
Into step 5B.
Knowledge is sentenced in step S5B, the current start-stop of driving
According to pre-stage test, train travel amplitude threshold A is setLVibrated when being and being travelled with 30km/h under train Light Condition
Vibration signal amplitude=400 that sensor measurement is obtained.
Signal analysis shows (Fig. 9), in moment 7.6s, the ＜ A of amplitude=390L, it is defined as end time T of driving a vehicle4=
7.6s.Train travel passes through 3.1s.
Into step S6.
Step S6, the current state of driving sentence knowledge
From moment value T32Rise along time shaft access time-Frequency point and to the amplitude of selected T/F point to the right
Carry out least square fitting, whenever fitting coefficient be less than 0 when count is incremented, and continue along time shaft move right fitting until
Moment value T4；Finally it is counted as N, compartment joint number=(N+1)/2.As a result show, at time 4.6s, fitting coefficient be -1.5,
Less than 0, begin to count 1.Until moment value t4Place, counts N=17, obtains compartment joint number=(N+1)/2=7 altogether.
Calculated according to compartment joint number 9 and single-unit railway car length 25m and obtain overall length of the train degree for 225m, it is total according to train
Length calculates with driving transit time and obtains train travel speed for 72.6km/h.As a result show, train by when normal operation.
Claims (9)
1. falling rocks judges recognition methods with train signal classification along railway, potential falling rocks generation site arrangement is shaken along railway
Dynamic sensor gathers, transmits primary signal to arithmetic center in real time, and arithmetic center obtains original time-amplitude signal and according to as follows
Step is implemented：
Step S1, denoising
Flip-flop, long period skew, the low-frequency anomaly in primary signal are filtered out, filtered signal is obtained, into step S2；
Step S2, I grades of vibration signal sentence knowledge
Amplitude threshold A is sets, as filtered signal amplitude >=AsWhen, it is I grades of vibration signals to sentence knowledge, records this moment value T1,
Into step S3；The amplitude threshold AsDrawn by falling rocks vibration test result and/or historical summary data statistics result
Falling rocks vibration amplitude variation characteristic is determined；
It is characterized in that：
Step S3, II grades of vibration signal sentence knowledge
Step S31, FFT
Filtered signal is subjected to FFT, obtained from moment value T1The frequency-amplitude signal risen, into step S32；
Step S32, signal sentence knowledge
Frequency-energy signal is obtained by frequency-amplitude signal of change obtained by step S31, energy ratio is calculated；The energy ratio
It is the ratio of signal energy value and whole signal energy values in 0Hz~500H frequency ranges in frequency-energy signal；
If energy ratio >=energy ratio threshold value RE, it is II grades of falling rocks vibration signals to sentence knowledge, records quarter time value T21, exit step
S32 enters step S4；If energy ratio ＜ energy ratio threshold values RE, it is II grades of driving vibration signals to sentence knowledge, is worth record quarter time
T22, exit step S32 enter step S4；
The energy ratio threshold value RE=80%~90%；
Step S4, III grade of vibration signal sentence knowledge
Step S41, STFT is converted
Filtered signal is subjected to STFT conversion, obtained from moment value T21Or T22T/F-the energy signal risen, into step
Rapid S42；
Step S42, signal sentence knowledge
From quarter time value T21Or T22Rise along time shaft access time-Frequency point to the right, and calculate at the time of chosen-Frequency point
Energy value average, when energy value average >=average energy value threshold value A0When be determined as vibration, continuation is chosen to the right along time shaft
T/F point, when the energy value average ＜ average energy value threshold value As in continuous 2s0When, it is judged as that vibration terminates, energy value is equal
It is worth threshold value A0=(1.5As)2；
Least square fitting is carried out to the energy value of the T/F point in the vibration beginning and ending time, during obtaining this time vibration
Fitting coefficient set；If fitting coefficient set meets energy fit threshold LSF1When, it is III grade of falling rocks vibration signal to sentence knowledge, it is determined that
Moment T occurs for falling rocks31It is energy value average >=average energy value threshold value A0At the time of；If fitting coefficient set meets energy fitting threshold
Value LSF2When, it is III grade of driving vibration signal to sentence knowledge, it is determined that driving is by occurring moment T32It is energy value average ＜ average energy values
Threshold value A0At the time of；
The energy fit threshold LSF1Be in fitting coefficient set 60% and the above fitting coefficient≤- 3, energy fitting
Threshold value LSF2Be in fitting coefficient set 60% and the above fitting coefficient ＞ -3；
Judge whether occur falling rocks according to III grade of falling rocks vibration signal identifying result, according to III grade of driving vibration signal identifying result
Judge whether it is train passage.
2. it is classified the falling rocks energy for judging that recognition methods is realized using falling rocks along the railway described in claim 1 and train signal
Scale computational methods, it is characterised in that：
In the step S42, after sentencing knowledge for III grade of falling rocks vibration signal, quarter time value T is recorded31, exit step S42, which enters, to be walked
Rapid S5A；
Step S5A, falling rocks energy scale computational methods；
Step S5A1, Energy Simulation calculating beginning and ending time sentence knowledge
Using STA/LTA methods to T1Filtered signal carries out Vibration identification from moment；When STA/LTA ratios >=STA/LTA ratios
Threshold value RSLWhen be defined as falling rocks Energy Simulation zero computing time t1, as amplitude≤amplitude lower threshold AminWhen be defined as falling rocks
Energy Simulation calculates end time t2；
The short cycle t of STA/LTAS, STA/LTA long periods tL, STA/LTA fractional thresholds RSLBy falling rocks vibration test result and/or
The falling rocks vibration amplitude variation characteristic determination that historical summary data statistics result is drawn, amplitude lower threshold AminShaken by falling rocks
The falling rocks vibration amplitude variation characteristic that dynamic test result and/or historical summary data statistics result are drawn is determined；
Step S5A2, calculating falling rocks scale
Falling rocks, which is calculated, according to formula 5.1 tumbles the scale A that releases energyE；
In formula, E (t) --- moment t1Square of~moment t RMS amplitude,
E0--- noise level, calculate and determine according to formula 5.2,
In formula, A1--- time t1Locate amplitude,
A2--- time t2Locate amplitude.
3. falling rocks energy scale computational methods according to claim 2, it is characterised in that：
The STA/LTA methods, tS=0.05s, tL=1s, RSL=3~3.8, window function ω [n] are that the Chinese that length is 128 is peaceful
Window, is calculated according to formula 5.3,5.4：
The amplitude lower threshold Amin=10%Amax~15%Amax, the AmaxIt is t1Amplitude is most in~moment t time interval
Big value.
Judge that the driving that recognition methods is realized is passed through 4. being classified using falling rocks along the railway described in claim 1 and train signal
Beginning and ending time sentences knowledge method, it is characterised in that：
In the step S42, after sentencing knowledge for III grade of driving vibration signal, quarter time value T is recorded32, exit step S42, which enters, to be walked
Knowledge is sentenced in rapid S5B, the current start-stop of driving；
The step S5B includes：As amplitude ＜ train travel amplitude thresholds ALWhen be defined as drive a vehicle end time T4；The train
Amplitude threshold of driving a vehicle ALIt is the vibration signal width that vibrating sensor measurement is obtained when being travelled with pre-set velocity under train Light Condition
Value.
5. the driving current beginning and ending time according to claim 4 sentences knowledge method, it is characterised in that：The pre-set velocity is
30km/h。
The state monitoring method 6. driving for sentencing the realization of knowledge method using the driving current beginning and ending time described in claim 5 is passed through,
It is characterized in that：It is determined that driving end time T4Sentence knowledge into step S6, the current state of driving afterwards；
The step S6 includes：From moment value T32Rise along time shaft access time-Frequency point and to selected T/F to the right
The amplitude of point carries out least square fitting, and count is incremented when fitting coefficient is less than 0, and continues to move right along time shaft
Dynamic fitting is until moment value T4；Finally it is counted as N, compartment joint number=(N+1)/2；
Overall length of the train degree is obtained according to compartment joint number and single-unit railway car length computation, it is current with driving according to overall length of the train degree
Time Calculation obtains train travel speed, and judges whether driving passage rate is normal.
7. according to any described method of claim 1~6, it is characterised in that：In the step S1, using the rank of 6 ranks~10
FIR is to original time-amplitude signal denoising.
8. according to any described method of claim 1~6, it is characterised in that：In the step S1, amplitude threshold AsIt is background
Signal amplitude Ab1.5 times, the background signal amplitude AbBe after vibrating sensor in-site installation under nature initial 1min it
Interior signal amplitude average value.
9. according to any described method of claim 1~6, it is characterised in that：The sensor be arranged on rail inner side or
On person's tie.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710065942.1A CN107063339B (en) | 2017-02-06 | 2017-02-06 | Falling rocks and train signal classification classification judge recognition methods along railway |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710065942.1A CN107063339B (en) | 2017-02-06 | 2017-02-06 | Falling rocks and train signal classification classification judge recognition methods along railway |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN107063339A true CN107063339A (en) | 2017-08-18 |
| CN107063339B CN107063339B (en) | 2019-04-16 |
Family
ID=59598775
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710065942.1A Active CN107063339B (en) | 2017-02-06 | 2017-02-06 | Falling rocks and train signal classification classification judge recognition methods along railway |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107063339B (en) |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112016180A (en) * | 2019-05-28 | 2020-12-01 | 浙江德盛铁路器材股份有限公司 | Method for analyzing quality condition of railway track basic equipment |
| CN112542046A (en) * | 2020-12-07 | 2021-03-23 | 无锡科晟光子科技有限公司 | Early warning monitoring method for long-distance pipeline heavy vehicle based on DAS |
| CN113661533A (en) * | 2019-04-04 | 2021-11-16 | 株式会社电装 | Falling object determination device, driving assistance system, and falling object determination method |
| CN113959553A (en) * | 2021-10-29 | 2022-01-21 | 中大检测(湖南)股份有限公司 | Intelligent low-frequency vibration sensor based on CPLD |
| US20220137595A1 (en) * | 2020-11-03 | 2022-05-05 | Verizon Patent And Licensing Inc. | Systems and methods for utilizing acoustic sensing to minimize a potential of damage to fiber optic cables |
| CN116182086A (en) * | 2023-04-28 | 2023-05-30 | 中特检管道工程（北京）有限公司 | Oil gas pipeline monitoring system adopting Internet of things technology |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JPH0488806A (en) * | 1990-07-30 | 1992-03-23 | Kawasaki Heavy Ind Ltd | Air-gap controller for linear motor |
| JPH0843148A (en) * | 1994-08-02 | 1996-02-16 | West Japan Railway Co | Landslide detector |
| US6216985B1 (en) * | 1997-08-29 | 2001-04-17 | Robert Douglas Stephens | Railway hazard acoustic sensing, locating, and alarm system |
| JP2004205309A (en) * | 2002-12-25 | 2004-07-22 | Ohbayashi Corp | Rock fall detecting system |
| CN102080985A (en) * | 2009-11-30 | 2011-06-01 | 清华大学 | Vibration measurement equipment for rail and measurement system using same |
| CN102079319A (en) * | 2009-11-30 | 2011-06-01 | 同方威视技术股份有限公司 | Railway rockfall early warning system and method |
| CN102354433A (en) * | 2011-10-14 | 2012-02-15 | 武汉理工光科股份有限公司 | Railway side slope crag rockfall monitoring and warning system based on fiber grating sensing technology |
-
2017
- 2017-02-06 CN CN201710065942.1A patent/CN107063339B/en active Active
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JPH0488806A (en) * | 1990-07-30 | 1992-03-23 | Kawasaki Heavy Ind Ltd | Air-gap controller for linear motor |
| JPH0843148A (en) * | 1994-08-02 | 1996-02-16 | West Japan Railway Co | Landslide detector |
| US6216985B1 (en) * | 1997-08-29 | 2001-04-17 | Robert Douglas Stephens | Railway hazard acoustic sensing, locating, and alarm system |
| JP2004205309A (en) * | 2002-12-25 | 2004-07-22 | Ohbayashi Corp | Rock fall detecting system |
| CN102080985A (en) * | 2009-11-30 | 2011-06-01 | 清华大学 | Vibration measurement equipment for rail and measurement system using same |
| CN102079319A (en) * | 2009-11-30 | 2011-06-01 | 同方威视技术股份有限公司 | Railway rockfall early warning system and method |
| CN102354433A (en) * | 2011-10-14 | 2012-02-15 | 武汉理工光科股份有限公司 | Railway side slope crag rockfall monitoring and warning system based on fiber grating sensing technology |
Cited By (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113661533A (en) * | 2019-04-04 | 2021-11-16 | 株式会社电装 | Falling object determination device, driving assistance system, and falling object determination method |
| CN112016180A (en) * | 2019-05-28 | 2020-12-01 | 浙江德盛铁路器材股份有限公司 | Method for analyzing 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 |
| US20220137595A1 (en) * | 2020-11-03 | 2022-05-05 | Verizon Patent And Licensing Inc. | Systems and methods for utilizing acoustic sensing to minimize a potential of damage to fiber optic cables |
| CN112542046A (en) * | 2020-12-07 | 2021-03-23 | 无锡科晟光子科技有限公司 | Early warning monitoring method for long-distance pipeline heavy vehicle based on DAS |
| CN112542046B (en) * | 2020-12-07 | 2022-03-11 | 无锡科晟光子科技有限公司 | Early warning monitoring method for long-distance pipeline heavy vehicle based on DAS |
| CN113959553A (en) * | 2021-10-29 | 2022-01-21 | 中大检测(湖南)股份有限公司 | Intelligent low-frequency vibration sensor based on CPLD |
| CN116182086A (en) * | 2023-04-28 | 2023-05-30 | 中特检管道工程（北京）有限公司 | Oil gas pipeline monitoring system adopting Internet of things technology |
| CN116182086B (en) * | 2023-04-28 | 2023-06-23 | 中特检管道工程（北京）有限公司 | Oil gas pipeline monitoring system adopting Internet of things technology |
Also Published As
| Publication number | Publication date |
|---|---|
| CN107063339B (en) | 2019-04-16 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107063339B (en) | Falling rocks and train signal classification classification judge recognition methods along railway | |
| CN106845429A (en) | Vibration signal classification judges recognition methods, falling rocks energy scale computational methods, falling rocks danger early warning method | |
| CN106274983B (en) | A kind of Mechanism Diagnosis method for identifying track traffic wheel On Wheel Rim Fracture | |
| CN105092467B (en) | A kind of quick monitoring device of high-speed railway wheel tread flat and method | |
| CN109100143A (en) | Fault Diagnosis of Roller Bearings and equipment based on CEEMDAN and CFSFDP | |
| CN206583854U (en) | A kind of monitoring device of the elastic state of rail fastening | |
| CN103473932B (en) | A kind of sound signal model recognition system of combination vibration graticule | |
| CN105489019B (en) | A kind of traffic throughput monitor system for dividing vehicle based on double-audio signal collection | |
| CN105928702A (en) | Variable working condition gear case bearing fault diagnosis method based on form component analysis | |
| CN110411766A (en) | The snakelike unstability detection method of train bogie, device, system and storage medium | |
| CN105424366A (en) | Bearing fault diagnosis method based on EEMD adaptive denoising | |
| CN105258789A (en) | Method and device for extracting vibration signal characteristic frequency band | |
| CN113358380B (en) | Rail vehicle snaking motion stability detection and evaluation method | |
| CN113343928B (en) | Method and device for detecting rail wave mill of speed change section high-speed railway and computer equipment | |
| CN103175897B (en) | A kind of high-speed switch hurt recognition methods based on vibration signal end-point detection | |
| CN108254179A (en) | A kind of bullet train wheel set bearing method for diagnosing faults based on MEEMD arrangement entropys | |
| CN105467428A (en) | Seismic wave warning method based on short-time energy detection and spectrum feature analysis | |
| CN104457643B (en) | A kind of impulse noise filter method and device of track geometry detection data | |
| CN109443769A (en) | Detection method, system, device and the readable storage medium storing program for executing of motor bearings state | |
| CN107423692A (en) | A kind of rail corrugation fault detection method based on wavelet-packet energy entropy | |
| CN115758289B (en) | Rail wave mill identification method based on multitask learning neural network | |
| CN109059839B (en) | Method, device and system for diagnosing out-of-roundness fault of wheel tread | |
| CN103194942B (en) | Track vibration signal motion detection device and detection method | |
| CN104949848A (en) | Hunting instability detecting algorithm based on frequency domain kurtosis | |
| CN113281414B (en) | Method and device for identifying short-wave disease types of steel rails and electronic equipment |
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 | ||
| TA01 | Transfer of patent application right | ||
| TA01 | Transfer of patent application right |
Effective date of registration: 20190118 Address after: 610041 No. 9, 4th Section of Renmin South Road, Wuhou District, Chengdu City, Sichuan Province Applicant after: Chengdu Mountainous Disasters and Environment Inst., Chinese Academy of Sciences Applicant after: Southwest Jiaotong University Address before: 610041 No. 9, 4th Section of Renmin South Road, Wuhou District, Chengdu City, Sichuan Province Applicant before: Chengdu Mountainous Disasters and Environment Inst., Chinese Academy of Sciences |
|
| GR01 | Patent grant | ||
| GR01 | Patent grant |