CN205262744U - Train wheel pair bearing trouble transient state characteristic detection device based on parameterization doppler transient state model - Google Patents

Abstract

The utility model discloses a train wheel pair bearing trouble transient state characteristic detection device based on parameterization doppler transient state model, the sound source signal, sensor, signal conditioning ware, data acquisition system and the central processing unit that send including the bearing, the sound source signal is received to the sensor, the sensor deliver a letter in the signal conditioning ware, the signal conditioning ware deliver a letter in data acquisition system, data acquisition system deliver a letter in central processing unit. This train wheel pair bearing trouble transient state characteristic detection device based on parameterization doppler transient state model can handle the train bearing signal that influenced by doppler effect, the accurate trouble of diagnosing the bearing.

Description

Railway wheelset bearing fault transient state characteristic checkout gear based on parametrization Doppler transient model

Technical field

The utility model relates to the analyzing and testing field of signal, is specifically related to a kind of based on parametrization Doppler winkThe Railway wheelset bearing fault transient state characteristic checkout gear of states model.

Background technology

Develop rapidly due to socioeconomic, to the demand of transportation technology also in continuous reinforcement. As a masterThe vehicles of wanting, train has very strong transport capacity and very high speed, is bringing into play very at societyLarge effect. But catastrophic failure may make train conveyance system produce serious accident, and bearing props upSupportting all wt of high-speed motion train, their fault is cause railway traffic instrument accident mainly formerCause, be therefore necessary very much to develop a kind of can be accurately and the key technology of automatic diagnosis bearing fault.

The detection of bearing fault signal relates to the processing to the signal in motion, and difficulty is large, is signal detectionThe one large difficult point in field. A large amount of methods is studied examines for the bearing fault being arranged in stationary machineDisconnected. Time frequency analysis is that a kind of effective extraction comprises the mode of non-stationary signal at interior mechanical health and fitness information, andAnd it can identification signal frequency component, discloses their time varying characteristic. As an adaptive decomposition method,Ensemble average empirical mode decomposition method (EnsembleEmpiricalModeDecomposition, EEMD)The signal of nonlinear and nonstationary can be resolved into one group of intrinsic mode letter according to their mode of oscillation of itselfNumber, has been widely used in bearing failure diagnosis field. In addition, in the diagnosis of bearing fault signal,Accidental resonance is also used as a kind of method that can utilize noise to increase output signal-to-noise ratio. Match tracingIt is the another kind of adaptive approach of selecting optimum atom rough estimate signal by iteration. But, due to manyGeneral existence of strangling effect, there will be frequency displacement in the signal of motion bearings, bandspreading and Modulation and Amplitude Modulation phenomenon.Generally speaking, said method all can not effectively address this problem.

Because above-mentioned defect, the design people, actively research and innovation in addition, a kind of based on ginseng to foundingThe Railway wheelset bearing fault transient state characteristic checkout gear of numberization Doppler transient model, makes it have more industryOn value.

Utility model content

For solving the problems of the technologies described above, the purpose of this utility model is to provide a kind of based on parametrization Doppler winkThe Railway wheelset bearing fault transient state characteristic checkout gear of states model, this device can be processed and be subject to Doppler effectThe train bearing signal of impact, the fault of Precise Diagnosis bearing.

A Railway wheelset bearing fault transient state characteristic checkout gear based on parametrization Doppler transient model,It is characterized in that: comprise sound-source signal, sensor, signal conditioner, data collecting system that bearing sendsAnd central processing unit, described sensor receives sound-source signal, and described sensor is messaging in described signal conditioner,Described signal conditioner is messaging in described data collecting system, and described data collecting system is messaging in described central authoritiesProcessor.

Further, described sensor is microphone.

Further, described data collecting system is DAS data collecting system.

Further, described central processing unit is computer.

By such scheme, the utility model at least has the following advantages: the utility model provides a kind of knotRailway wheelset bearing fault transient state characteristic checkout gear based on parametrization Doppler transient model can detect realityThe related data of border motion bearings fault-signal, by Doppler's model and the actual motion bearing fault set upRelevance parameter between signal is optimized. Finally, motion bearings fault can be passed through optimum Doppler's winkThe initial period model parameter that state relevant matches model is corresponding is diagnosed. Be subject to Doppler effect thereby can processThe train bearing signal of impact, the fault of Precise Diagnosis bearing.

Above-mentioned explanation is only the general introduction of technical solutions of the utility model, new in order to better understand this practicalityThe technological means of type, and can being implemented according to the content of description, below with better reality of the present utility modelExecute example and coordinate accompanying drawing to be described in detail as follows.

Brief description of the drawings

Fig. 1 is a kind of Railway wheelset bearing fault wink based on parametrization Doppler transient model of the utility modelThe structural representation figure of state feature detection device;

Fig. 2 is the schematic diagram of Doppler effect;

Fig. 3 is the signal time-domain diagram collecting under the malfunction of the utility model embodiment middle (center) bearing outer ring;

Fig. 4 is the spectrogram collecting under the malfunction of the utility model embodiment middle (center) bearing outer ring;

Fig. 5 is under the malfunction of the utility model embodiment middle (center) bearing outer ring, builds according to the utility model methodVertical Doppler's transient state relevant matches illustraton of model;

Fig. 6 is the detection figure of train bearing outer ring fault-signal in the utility model embodiment;

Fig. 7 is under the malfunction of the utility model embodiment middle (center) bearing outer ring, builds according to the utility model methodThe vertical period transient state illustraton of model based on Laplace small echo relevant to Doppler's transient state relevant matches model.

Detailed description of the invention

For making utility model object of the present utility model, feature, advantage can be more obvious and understandable,Below in conjunction with the accompanying drawing in the utility model embodiment, the technical scheme in the utility model embodiment is enteredRow is described clearly and completely, and obviously, the embodiments described below are only that the utility model part is realExecute example, but not whole embodiment. Based on the embodiment in the utility model, those of ordinary skill in the artDo not making the every other embodiment obtaining under creative work prerequisite, all belonging to the utility model and protectThe scope of protecting.

In order to understand better the utility model numerical procedure, below with outer ring fault detect as an example, rightThe application of the described Railway wheelset bearing fault transient state characteristic checkout gear based on parametrization Doppler transient modelTell about in detail:

Embodiment: a kind of Railway wheelset bearing fault transient state characteristic inspection based on parametrization Doppler transient modelSurvey device, comprise sound-source signal 1, sensor 2, signal conditioner 3, data collecting system that bearing sends4 and central processing unit 5, described sensor receives sound-source signal, and described sensor is messaging in described signal conditionDevice, described signal conditioner is messaging in described data collecting system, described in described data collecting system is messaging inCentral processing unit.

Described sensor is microphone.

Described data collecting system is DAS data collecting system.

Described central processing unit is computer.

While breaking down in train bearing outer ring in motion, due to the impact of Doppler effect, can cause detectingSignal is modulated, the available Railway wheelset axle based on parametrization Doppler transient model described in the utility modelHolding fault transient state characteristic checkout gear detects.

The train bearing signal detecting is y (t), Doppler's transient model parameter of foundation is optimized, whenWhen Doppler's transient model and physical fault signal reach maximum correlation coefficient, can be by optimum DopplerThe cycle parameter of the period transient state model that transient model is corresponding and the motion bearings fault signature cycle phase calculatingRelatively, draw fault type.

Concrete mode of operation is as follows:

The parameterized model of model based on monolateral Laplace small echo, by the theoretical value note of bearing fault signalForDescribed bearing fault signal can be expressed as:

Wherein:For the theoretical value of bearing fault signal, U is the time span of signal, and τ prolongsLate the time, ζ is damped coefficient, and f is frequency. Under note τ, ζ and f, scope is T_d, Z and F, so:

$T_d,F\⋐R^{+}$

$Z\⋐(\[0,1)\∩R^{+})---\left(2\right)$

In formula (1)T, τ, ζ and f all represent variable;

Then set up one-period model by introducing parameter T, simulate the ripple of bearing fault signal with thisShape feature, can constructed fuction:

Revise again the periodic model based on monolateral Laplace small echo, by sensor receive acoustical signal time engraveFor { t_R, accept the moment can be expressed as:

{t_R}＝{t₀,t₀+1/f_s,t₀+2/f_s,t₀+(N-1)/f_s}(4)

Wherein, fs is frequency, t₀Be the initial time of sound-source signal, N is data length;

By position relationship, t_RCan also be expressed as:

$t_R=t_e+\mathrm{\Δ}t=t_e+R/V_{sw}=t_e+\sqrt{r^2+{(S-L)}^2}/V_{sw}---\left(5\right)$

Wherein, R is the distance between sound source and sensor, V_swAirborne velocity of sound, t_eSending out of acoustical signalGo out the moment, r is the distance between sensor and sound source traffic direction. L is the immediate movement of sound source.

For the L in formula (5), can also be obtained by following formula:

$L\left(t\right)={\∫}_0^t{V}_{s}dt---\left(6\right)$

ψ in formula (3)_periodic(t) t in is the t in formula (5)_e，ψ_periodic(t) can change { ψ into_e(t_e)}；

Periodic model based on monolateral Laplace small echo is applied to Doppler effect, show that corresponding parametrization is manyThe general transient model of strangling, in the time that acoustical signal propagates into receiver from sound source, the acoustical signal receiving is modulated,By Morse acoustic theory, (M=V in the time that sound source moves with subsonic_s/V_sw< 0.2), just giving tacit consent to sound source is oneIndividual first order pole sound source, the acoustical signal receiving is expressed as:

$P=\frac{q^{\&prime;}\&lsqb;t-(R/V_{sw})\&rsqb;}{4\mathrm{\&pi;}R{(1-Mcos\mathrm{\&theta;})}^2}+\frac{q\&lsqb;t-(R/V_{sw})\&rsqb;\&lsqb;cos\mathrm{\&theta;}-M\&rsqb;V_s}{4{\mathrm{\&pi;R}}^2{(1-Mcos\mathrm{\&theta;})}^3}=P_A+P_B---\left(7\right)$

Wherein q is total mass flow rate, and t is the propagation time of sound, M=V_s/V_swIt is the Mach number of sound source speed.θ is the angle that sound source traffic direction and sound source arrive sensor place straight line, in formula (7), and P_AExpression connectsThe acoustic pressure P receiving is inversely proportional to parameters R, P_BRepresent near-field effect. In the time of M < 0.2, P_BJust can be byIgnore, therefore, the acoustic pressure receiving can be expressed as:

$P\&ap;\frac{q^{\&prime;}\&lsqb;t-(R/V_{sw})\&rsqb;}{4\mathrm{\&pi;}R{(1-Mcos\mathrm{\&theta;})}^2}---\left(8\right)$

Also can be write as:

$P\&ap;\frac{r}{{(1-Mcos\mathrm{\&theta;})}^{2}R}\&CenterDot;\frac{\&lsqb;t-(R/V_{sw})\&rsqb;q^{\&prime;}}{4\mathrm{\&pi;}r}---\left(9\right)$

Wherein, r/ (R (1-Mcos θ)²) play amplitude-modulated effect, q ' [t-(R/V_sw)]/(4 π are r) sound sources and connectThe acoustic pressure receiving while not relatively moving between receipts device.

The signal receiving can be written as:

${\mathrm{\&psi;}}_{\mathrm{Re}c}={\mathrm{\&psi;}}_e\&CenterDot;\frac{r/R}{{(1-M\cos \mathrm{\&theta;})}^2}---\left(10\right)$

The coefficient correlation of calculating different parameters drag and actual signal specifically comprises: will be designated as coefficient correlation,Coefficient correlation can be expressed as:

${\mathrm{\&eta;}}_{a\left(n\right),b\left(n\right)}=\frac{<a\left(n\right),b\left(n\right)>}{\sqrt{<a\left(n\right),a\left(n\right)>}\sqrt{<b\left(n\right),b\left(n\right)>}}=\frac{\underset{i=1}{\overset{Q}{\&Sigma;}}a\left(i\right)\&times;b\left(i\right)}{\sqrt{\underset{i=1}{\overset{Q}{\&Sigma;}}a\left(i\right)\&times;a\left(i\right)}\sqrt{\underset{i=1}{\overset{Q}{\&Sigma;}}b\left(i\right)\&times;b\left(i\right)}}---\left(11\right)$

Wherein a (n), b (n) is for there being two groups of data of equal length, and n is their data length, η_a(n),b(n)Be their coefficient correlation,<a (n), b (n)>account form be:

$<a\left(n\right),b\left(n\right)>=\underset{i=1}{\overset{Q}{\&Sigma;}}a\left(i\right)\&times;b\left(i\right)---\left(12\right)$

Coefficient correlation η_a(n),b(n)Codomain be:

-1≤η_a(n),b(n)≤1(13)

And work as η_a(n),b(n)Close to 0 o'clock, can think a (n), b (n) is linear correlation.

By different cycles parameter T, Doppler's transient model that damped coefficient ζ and discrete frequency f set up and trueThe coefficient correlation of train bearing fault-signal is removed optimization cycle transient model and many as a quantitative meansThe general transient model of strangling;

The formula of correlation analysis can be expressed as:

${\mathrm{\&rho;}}_{y\left(t\right),{\mathrm{\&psi;}}_{\mathrm{Re}c}\left(t\right)}=\frac{<y\left(t\right),{\mathrm{\&psi;}}_{\mathrm{Re}c}\left(t\right)>}{\sqrt{<y\left(t\right),y\left(t\right)>}\sqrt{<{\mathrm{\&psi;}}_{\mathrm{Re}c}\left(t\right),{\mathrm{\&psi;}}_{\mathrm{Re}c}\left(t\right)>}}=\frac{\underset{t=t_0}{\overset{t_N}{\&Sigma;}}y\left(t\right)\&times;{\mathrm{\&psi;}}_{\mathrm{Re}c}\left(t\right)}{\sqrt{\underset{t=t_0}{\overset{t_N}{\&Sigma;}}y\left(t\right)\&times;y\left(t\right)}\sqrt{\underset{t=t_0}{\overset{t_N}{\&Sigma;}}{\mathrm{\&psi;}}_{\mathrm{Re}c}\&times;{\mathrm{\&psi;}}_{\mathrm{Re}c}\left(t\right)}}---\left(14\right)$

Wherein, y (t) is the amplitude of train bearing fault-signal;

In optimized process, when the detection of parametrization Doppler transient model and train bearing fault-signalWhen value has maximum correlation coefficient, just think optimization of model, think Doppler's transient state nowThe period transient state model that model is corresponding has disclosed real train bearing fault transient components.

Experiment one: experimental subjects is mainly fault bearing. The running parameter of bearing can reference table 1, all the other experimentsParameter can reference table 2. Testing failure bearing NJ (P) 3226XI profile shaft holds, and bearing outer ring surface has oneWhat 0.18mm was wide runs through slight crack fault.

Table 1 bearing working parameter

All the other parameters of using in table 2 first experiment

Experiment two is to implement according to Doppler's theoretical model. Experiment parameter is Vsw=340m/s, r=2m, Vs=30m/s, S=3.5m. Sound letter under recording play first experiment on the automobile of operation inNumber, the sensor in roadside receives signal. Experiment two and experiment one use identical sensor and DAS.

In theory, in the time adopting the parameter of table one and table two, the fault signature of outer ring is 138.74Hz, therefore,The periodic shock sigtnal interval that outer ring fault is relevant is 0.007s.

Fig. 3 and Fig. 4 are respectively time-domain diagram and the spectrogram of the detected value of outer ring fault-signal, can find out,Under the impact of Doppler effect, fault-signal generation frequency displacement, amplitude is also modulated. By traditional sideFormula cannot draw failure-frequency from the information of distortion.

As shown in Fig. 5-7, one of the model parametrization periodic model based on Laplace small echo, then willDoppler effect is added on model, obtains Doppler's transient model, finally according to correlation analysis, optimizes ginsengNumber, in the time that the transient state interval of periodic model is selected as 0.007s, Doppler's transient model and reality that it is correspondingThe coefficient correlation of border fault-signal reaches maximum, now, the cycle parameter of optimal period transient model andReal bearing outer ring fault is impacted interval and is met. Result shows, based on relevant of parametrization Doppler transient stateJoin model and accurately shown that fault phase Guan pulse in the fault-signal of outer ring is washed into point.

Can find out from analytic process and application example, the utility model provide based on parametrization Doppler winkThe Railway wheelset bearing fault transient state characteristic checkout gear of states model, can be subject to Doppler at train bearing signalUnder effects, the relevant transient pulse composition of its fault is effectively identified, be out of order thereby diagnose, its spyPoint has determined that the method can effectively be applied to Railway wheelset bearing failure diagnosis.

The above is only preferred embodiment of the present utility model, is not limited to the utility model, shouldWhen pointing out, for those skilled in the art, do not departing from the utility model know-whyPrerequisite under, can also make some improvement and modification, these improve and modification also should be considered as the utility modelProtection domain.

Claims (4)

1. the Railway wheelset bearing fault transient state characteristic based on parametrization Doppler transient model detects dressPut, it is characterized in that: comprise sound-source signal, sensor, signal conditioner, data acquisition that bearing sendsSystem and central processing unit, described sensor receives sound-source signal, and described sensor is messaging in described signal and adjustsReason device, described signal conditioner is messaging in described data collecting system, and described data collecting system is messaging in instituteState central processing unit.

2. the Railway wheelset bearing fault based on parametrization Doppler transient model according to claim 1Transient state characteristic checkout gear, is characterized in that: described sensor is microphone.

3. the Railway wheelset bearing fault based on parametrization Doppler transient model according to claim 2Transient state characteristic checkout gear, is characterized in that: described data collecting system is DAS data collecting system.

4. the Railway wheelset bearing fault based on parametrization Doppler transient model according to claim 3Transient state characteristic checkout gear, is characterized in that: described central processing unit is computer.