CN109827776A - Bearing Fault Detection Method and system - Google Patents
Bearing Fault Detection Method and system Download PDFInfo
- Publication number
- CN109827776A CN109827776A CN201910202408.XA CN201910202408A CN109827776A CN 109827776 A CN109827776 A CN 109827776A CN 201910202408 A CN201910202408 A CN 201910202408A CN 109827776 A CN109827776 A CN 109827776A
- Authority
- CN
- China
- Prior art keywords
- module
- signal
- parameter
- resonance
- data
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 238000001514 detection method Methods 0.000 title claims abstract description 23
- 238000000354 decomposition reaction Methods 0.000 claims abstract description 58
- 230000008878 coupling Effects 0.000 claims abstract description 55
- 238000010168 coupling process Methods 0.000 claims abstract description 55
- 238000005859 coupling reaction Methods 0.000 claims abstract description 55
- 239000002245 particle Substances 0.000 claims abstract description 48
- 238000012545 processing Methods 0.000 claims abstract description 24
- 238000000034 method Methods 0.000 claims description 22
- 238000012216 screening Methods 0.000 claims description 17
- 238000000605 extraction Methods 0.000 claims description 6
- 230000000694 effects Effects 0.000 abstract description 8
- 238000005516 engineering process Methods 0.000 abstract description 5
- 238000010586 diagram Methods 0.000 description 25
- 230000003044 adaptive effect Effects 0.000 description 6
- 230000002708 enhancing effect Effects 0.000 description 5
- 238000005457 optimization Methods 0.000 description 5
- 238000005070 sampling Methods 0.000 description 5
- 238000004364 calculation method Methods 0.000 description 4
- 230000008569 process Effects 0.000 description 4
- 241000208340 Araliaceae Species 0.000 description 3
- 235000005035 Panax pseudoginseng ssp. pseudoginseng Nutrition 0.000 description 3
- 235000003140 Panax quinquefolius Nutrition 0.000 description 3
- 230000009977 dual effect Effects 0.000 description 3
- 235000008434 ginseng Nutrition 0.000 description 3
- 238000001228 spectrum Methods 0.000 description 3
- 230000008859 change Effects 0.000 description 2
- 238000002474 experimental method Methods 0.000 description 2
- 238000012544 monitoring process Methods 0.000 description 2
- 238000005096 rolling process Methods 0.000 description 2
- 230000003595 spectral effect Effects 0.000 description 2
- 238000010183 spectrum analysis Methods 0.000 description 2
- 238000012952 Resampling Methods 0.000 description 1
- 230000006978 adaptation Effects 0.000 description 1
- 238000013329 compounding Methods 0.000 description 1
- 238000001914 filtration Methods 0.000 description 1
- 230000006698 induction Effects 0.000 description 1
- 238000007689 inspection Methods 0.000 description 1
- 238000000608 laser ablation Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 210000003739 neck Anatomy 0.000 description 1
- 230000010355 oscillation Effects 0.000 description 1
- 230000000737 periodic effect Effects 0.000 description 1
- 230000003014 reinforcing effect Effects 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
- 238000012795 verification Methods 0.000 description 1
Landscapes
- Measurement Of Mechanical Vibrations Or Ultrasonic Waves (AREA)
Abstract
Bearing Fault Detection Method and system, comprising: acquisition original vibration signal calculates the original vibration signal with particle swarm algorithm and obtains resolution parameter；Optimize the resolution parameter, to obtain parameter combination information, decomposing the original vibration signal accordingly is Mode Decomposition information；Extract the coupling input signal in the Mode Decomposition information；Optimize bistable-state random resonance system with the particle swarm algorithm, accidental resonance processing is carried out to the coupling input signal accordingly, to obtain fault characteristic signals.It is poor that the present invention solves signal processing effect of the existing technology, the low technical problem of Weak fault signal detection precision.
Description
Technical field
The present invention relates to a kind of mechanical fault signals detection methods, more particularly to a kind of Bearing Fault Detection Method and are
System.
Background technique
Stochastic Resonance Signal Enhancement Method is to convert the random noise energy in signal to by the theory of nonlinear oscillation
The energy of signal specific, to achieve the purpose that enhance the signal strength, effectively be detected.
Variation mode decomposition is based on these three concepts of classical Wiener filtering, Hilbert transform and frequency compounding
Adaptive signal decomposition method.Compared to recurrence ' screening ' mode that empirical mode decomposition algorithm and local mean value decompose, become mode
Signal decomposition is converted the mode of onrecurrent, Variational Decomposition by decomposition algorithm, is substantially multiple adaptive wiener filter groups.No
It is same as empirical mode decomposition method, in becoming mode decomposition, intrinsic mode function natural mode of vibration component is no longer limited to narrowband letter
Number.But it needs that the parameter for becoming mode decomposition is previously set when becoming mode decomposition algorithm process signal, and uses mode decomposition merely
Algorithm can not accurately extract Weak fault information under strong noise environment, so that mode decomposition algorithm exists centainly in practical applications
Limitation.
In conclusion the fault-signal Enhancement Method of the prior art is there are signal processing effect is poor, the inspection of Weak fault signal
Survey the low technical problem of precision.
Summary of the invention
In view of the above prior art there are signal processing effect is poor, the low technical problem of Weak fault signal detection precision,
The purpose of the present invention is to provide a kind of Bearing Fault Detection Method and system, a kind of Bearing Fault Detection Method, comprising: acquisition
Original vibration signal calculates original vibration signal with particle swarm algorithm and obtains resolution parameter；Optimal Decomposition parameter, to obtain parameter group
Information is closed, decomposing original vibration signal accordingly is Mode Decomposition information；Extract the coupling input signal in Mode Decomposition information；With
Particle swarm algorithm optimizes bistable-state random resonance system, carries out accidental resonance processing to coupling input signal accordingly, to obtain event
Hinder characteristic signal.
In one embodiment of the present invention, calculate resolution parameter the step of, comprising: with vibration inductor obtain axis
The original vibration signal held；Particle iterative data is obtained according to original vibration signal, calculates fitness data accordingly；According to adaptation
Spend data setting population iterative logical；Resolution parameter is calculated to obtain according to population iterative logical.
In one embodiment of the present invention, decompose original vibration signal the step of, comprising: obtain envelope data, accordingly
Resolution parameter is calculated, to obtain parameter combination information；According to each parameter combination information setting mode decomposition logic；According to mode point
It solves logic and original vibration signal is decomposed into multiple modal components.
In one embodiment of the present invention, extract coupling input signal the step of, comprising: obtaining mode decomposed information
Kurtosis data；Available modal components are obtained according to kurtosis data screening Mode Decomposition information；Processing can be coupling with modal components
Input signal.
In one embodiment of the present invention, the step of handling signal of resonating, comprising: extract the coupling of each particle in population
System parameter is closed, calculates output signal-to-noise ratio accordingly；Speed and the position of particle are set according to output signal-to-noise ratio, calculate resonance accordingly
System update data；Bistable-state random resonance system is updated according to resonance system more new data；Accidental resonance handles coupling input
Signal acquisition fault characteristic signals.
In one embodiment of the present invention, a kind of bearing fault detection system, comprising: resolution parameter module, to adopt
Collect original vibration signal, original vibration signal is calculated with particle swarm algorithm and obtains resolution parameter；Decomposing module is joined to Optimal Decomposition
Number, to obtain parameter combination information, decomposing original vibration signal accordingly is Mode Decomposition information, decomposing module and resolution parameter mould
Block connection；Coupling input module, to extract the coupling input signal in Mode Decomposition information, coupling input module and decomposition mould
Block connection；Stable state resonance module, for optimizing bistable-state random resonance system with particle swarm algorithm, accordingly to coupling input signal
Accidental resonance processing is carried out, to obtain fault characteristic signals, stable state resonance module is connect with coupling input module.
In one embodiment of the present invention, resolution parameter module, comprising: original signal module, for vibration induction
The original vibration signal of device acquisition bearing；Fitness module, to obtain particle iterative data according to original vibration signal, according to
To calculate fitness data, fitness module is connect with original signal module；Iteration setting module, to according to fitness data
Population iterative logical is set, iteration setting module is connect with fitness module；Parameter calculating module, to be changed according to population
Resolution parameter is obtained for logic calculation, parameter calculating module is connect with iteration setting module.
In one embodiment of the present invention, decomposing module includes: parameter component module, to obtain envelope data, according to
To calculate resolution parameter, to obtain parameter combination information；Logic module is decomposed, to according to each parameter combination information setting mode
Logic is decomposed, logic module is decomposed and is connect with parameter component module；Component obtains module, to will be former according to mode decomposition logic
It is multiple modal components that beginning vibration signal, which decomposes, and component obtains module and connect with logic module is decomposed.
In one embodiment of the present invention, coupling input module, comprising: kurtosis module is decomposed to obtaining mode and believed
The kurtosis data of breath；Component screening module, to obtain available modal components according to kurtosis data screening Mode Decomposition information, point
Amount screening module is connect with kurtosis module；Component processing module can be coupling input signal, component with modal components to handle
Processing module is connect with component screening module.
In one embodiment of the present invention, stable state resonance module, comprising: SNR module, to extract in population
The coupled system parameter of each particle, calculates output signal-to-noise ratio accordingly；Data module is updated, grain is arranged according to output signal-to-noise ratio
The speed of son and position calculate resonance system more new data accordingly, update data module and connect with SNR module；Resonance updates
Module, to update bistable-state random resonance system, resonance update module and more new data mould according to resonance system more new data
Block connection；Feature obtains module, handles coupling input signal acquisition fault characteristic signals to accidental resonance, feature obtains module
It is connect with resonance update module.
As described above, a kind of Bearing Fault Detection Method provided by the invention and system have the advantages that the party
Method can reduce traditional VMD (Variational Mode Decomposition) decomposition algorithm vulnerable to penalty alpha, point
It measures number K and Lagrange multiplier and updates the influence of step-length tau, and the signal-to-noise ratio of output signal can be significantly increased, effectively mention
Take Weak fault signal characteristic.
To sum up, the present invention provides a kind of Bearing Fault Detection Method and system, solves at signal of the existing technology
It is poor to manage effect, the low technical problem of Weak fault signal detection precision.
Detailed description of the invention
Fig. 1 is shown as Bearing Fault Detection Method step schematic diagram of the invention.
Fig. 2 is shown as the specific flow chart of step S1 in one embodiment in Fig. 1.
Fig. 3 is shown as the specific flow chart of step S2 in one embodiment in Fig. 1.
Fig. 4 is shown as the specific flow chart of step S3 in one embodiment in Fig. 1.
Fig. 5 is shown as the specific flow chart of step S4 in one embodiment in Fig. 1.
Fig. 6 is shown as bearing fault monitoring system module schematic diagram of the invention.
Fig. 7 bearing fault detection system data flows to schematic diagram.
Fig. 8 is shown as the specific module diagram of resolution parameter module in one embodiment in Fig. 6.
Fig. 9 is shown as original signal time domain waveform and spectrogram of the invention.
Figure 10 is shown as particle optimizing envelope Entropy Changes schematic diagram of the invention.
Figure 11 is shown as the specific module diagram of decomposing module in one embodiment in Fig. 6.
Figure 12 is shown as intrinsic mode function time domain waveform of the invention.
Figure 13 is shown as envelope signal processing waveform diagram of the invention.
Figure 14 is shown as the specific module diagram of stable state resonance module in one embodiment in Fig. 6.
Figure 15 is shown as controlled system output signal-to-noise ratio value variation diagram of the invention.
Figure 16 is shown as feature enhancing time domain plethysmographic signal schematic diagram of the invention.
Component label instructions
1 resolution parameter module
2 decomposing modules
3 coupling input modules
4 stable state resonance modules
11 original signal modules
12 fitness modules
13 iteration setting modules
14 parameter calculating modules
21 parameter component modules
22 decompose logic module
23 components obtain module
31 kurtosis modules
32 component screening modules
33 component processing modules
41 SNR modules
42 update data module
43 resonance update modules
44 features obtain module
Step numbers explanation
S1~S4 method and step
S11~S14 method and step
S21~S23 method and step
S31~S33 method and step
S41~S44 method and step
Specific embodiment
Embodiments of the present invention are illustrated by particular specific embodiment below, those skilled in the art can be by this explanation
Content disclosed by book is understood other advantages and efficacy of the present invention easily.
Fig. 1 is please referred to Figure 16, it should however be clear that this specification structure depicted in this specification institute accompanying drawings, only to cooperate specification
Revealed content is not intended to limit the invention enforceable restriction item so that those skilled in the art understands and reads
Part, therefore do not have technical essential meaning, the modification of any structure, the change of proportionate relationship or the adjustment of size are not influencing
Under the effect of present invention can be generated and the purpose that can reach, should all still fall in disclosed technology contents can contain
In the range of lid.Meanwhile in this specification it is cited such as " on ", " under ", " left side ", " right side ", " centre " and " one " term,
It is merely convenient to being illustrated for narration, rather than to limit the scope of the invention, relativeness is altered or modified,
It is changed under technology contents without essence, when being also considered as the enforceable scope of the present invention.
Referring to Fig. 1, Bearing Fault Detection Method step schematic diagram of the invention is shown as, as shown in Figure 1, of the invention
It is designed to provide a kind of Bearing Fault Detection Method and system, a kind of Bearing Fault Detection Method, comprising:
S1, acquisition original vibration signal, calculate original vibration signal with particle swarm algorithm and obtain resolution parameter, in an embodiment
In, it can be first using minimum envelop entropy as fitness function；
S2, Optimal Decomposition parameter decompose original vibration signal accordingly to obtain parameter combination information as Mode Decomposition letter
Breath, in one embodiment, using particle swarm algorithm Optimizing Search VMD, (Variational Mode Decomposition becomes mould
Formula is decomposed) resolution parameter；
S3, the coupling input signal extracted in Mode Decomposition information are chosen Optimizing Search and are obtained most in one embodiment
Good parameter combination to bearing Weak fault vibration signal collected carry out VMD decomposition, obtain several intrinsic modal components (IMF,
Intrinsic Mode Function)；
S4, bistable-state random resonance system is optimized with particle swarm algorithm, accidental resonance is carried out to coupling input signal accordingly
Processing, to obtain fault characteristic signals, in one embodiment, selects the maximum intrinsic modal components of kurtosis by kurtosis criterion；
The matched coupled bistable accidental resonance of parameter adaptive finally is carried out to the IMF component, is realized to bearing Weak fault signal
Feature extraction.
Referring to Fig. 2, the step schematic diagram of step S1 in one embodiment in Fig. 1 is shown as, as shown in Fig. 2, calculating
The step S1 of resolution parameter, comprising:
S11, the original vibration signal that bearing is obtained with vibration inductor, in one embodiment, minimum envelop entropy is determined
Justice are as follows:
min{EP}={ EP1, EP2..., EPK}
K is the quantity of IMF, E in formulaPKRepresent the envelope spectrum entropy of k-th component.Wherein envelope spectrum entropy is defined as:
A (j) is the envelope signal that the original vibration signal acquired obtains after Hilbert envelope demodulates, P in formulajIt is a
(j) normalized form.；
S12, particle iterative data is obtained according to original vibration signal, calculates fitness data accordingly, in one embodiment,
Particle populations initialization, setting population quantity N, penalty alpha, component number K, Lagrange multiplier update step-length tau
And maximum number of iterations MAXIter.The search range that the search range of alpha is 100~2500, K is searching for 2~10, tau
Rope range is 0~1；
S13, according to fitness data setting population iterative logical；
S14, resolution parameter is calculated to obtain according to population iterative logical, in one embodiment, alpha is to punish in VMD parameter
Penalty factor optimal value, K are Decomposition order optimal value in VMD parameter, and tau is that Lagrange multiplier updates step-length most in VMD parameter
The figure of merit.
Referring to Fig. 3, the step schematic diagram of step S2 in one embodiment in Fig. 1 is shown as, as shown in figure 3, decomposing former
The step S2 of beginning vibration signal, comprising:
S21, envelope data is obtained, calculates resolution parameter accordingly, to obtain parameter combination information, in one embodiment, adopts
It is random initial in the range of setting in having been carried out example with the best VMD resolution parameter of particle swarm optimization algorithm Adaptive matching
Change population initial position and speed, using every group of parameter in initial population respectively to collect vibration signal carry out VMD at
Reason；
S22, according to each parameter combination information setting mode decomposition logic, in one embodiment, it is right to calculate every group of parameter institute
The envelope entropy of the intrinsic mode function IMF answered, using IMF component corresponding to minimum envelop entropy as the corresponding set of ginseng of use
Number carries out VMD treated optimal component；N number of minimum envelop entropy is obtained after the completion of every generation traversal, by population all particles
Minimum value is made comparisons with the global optimum of record in minimum envelop entropy, stores its smaller value as contemporary population global optimum
With corresponding parameter combination (alpha, K, tau)；
S23, original vibration signal is decomposed by multiple modal components according to mode decomposition logic, in one embodiment, obtained
To several intrinsic mode functions of N group IMF.
Referring to Fig. 4, the step schematic diagram of step S3 in one embodiment in Fig. 1 is shown as, as shown in figure 4, extracting coupling
Close the step S3 of input signal, comprising:
The kurtosis data of S31, obtaining mode decomposed information select maximum kurtosis value in one embodiment from K IMF
To be entered signal of the corresponding IMF component as dual input coupling bi-stable stochastic resonance theory；
S32, available modal components are obtained according to kurtosis data screening Mode Decomposition information；
S33, processing can be coupling input signal with modal components.
Referring to Fig. 5, the step schematic diagram of step S4 in one embodiment in Fig. 1 is shown as, as shown in figure 5, at resonance
Manage the step S4 of signal, comprising:
S41, the coupled system parameter for extracting each particle in population, calculate output signal-to-noise ratio accordingly, in one embodiment,
Using output signal-to-noise ratio as fitness value using the system ginseng of particle swarm optimization algorithm Adaptive matching coupling bi-stable stochastic resonance theory
The several and coefficient of coup is chosen in particle group optimizing search process corresponding to global maximum output signal-to-noise ratio in one embodiment
Parameter combination (a, b, r) carries out coupling bi-stable stochastic resonance theory signal enhancing to input signal after double sampling, to enhanced letter
Number restore former sample frequency and carry out spectrum analysis to update speed and the position of all particles according to the following formula in one embodiment:
vij(t+1)=wvij(t)+c1rand[Pbesti(t)-xij(t)]+c2rand[Nbest(t)-xij(t)]
xij(t+1)=xij(t)+vij(t+1)
In formula, i=1,2 ..., N, j=1,2 ..., m, m are population dimension；T is current iteration number；C1, c2 are study
The factor usually takes c1=c2=2；vijIt (t) is the current value of j-th of element in the velocity vector of i-th of particle, vij(t+1) it is
The updated value of j-th of element in i-th of particle rapidity；xij(t) it is updated for j-th of element in the position vector of i-th of particle
Value afterwards, xijIt (t+1) is the updated value of j-th of element in the position vector of i-th of particle；W is inertia weight, that is, is kept
The originally coefficient of speed.
The calculation formula of inertia weight is as follows:
Wmax, wmin are respectively inertia weight bound in formula, and t is current iteration number, and MAXIter is the maximum of setting
The number of iterations
S42, speed and the position that particle is arranged according to output signal-to-noise ratio, calculate resonance system more new data, one accordingly
In embodiment, one suitable mutative scale ginseng is chosen using method for resampling for the condition for meeting stochastic resonance system small parameter
Number k will be input to coupled system after gained coupling input signal double sampling in preceding step so that double sampling frequency is much smaller than 1
In, in one embodiment, judge whether to meet iteration stopping condition.Stop changing when updating algebra and reaching setting value MAXIter
Generation, and optimized parameter is exported, otherwise go to continuation iteration.Population at individual number is bigger in practical application, the number of iterations is bigger, sampling
Rate is bigger, and required calculation amount is also bigger, therefore MAXIter, N should rationally be arranged；Comprehensively considering calculation amount and population multiplicity
Under the premise of property, MAXIter generally takes 10~20, N generally to take 30~50.Wherein, output signal-to-noise ratio formula are as follows:
Ad is the amplitude at target frequency in formula, and N is sampling number, and Ai is in system output spectrum figure at every spectral line
Amplitude；
S43, bistable-state random resonance system is updated according to resonance system more new data, in one embodiment, couples bistable
Stochastic resonance system model are as follows:
S (t) is input periodic signal in formula, and x (t) is controlled system output, and y (t) is control system output, system parameter
The bistable system that a0, b0 take fixed value to be 1 is by control system, and the conduct control system of another Parameter adjustable, r are
The coefficient of coup, a, b are adjustable system parameter；
S44, accidental resonance handle coupling input signal acquisition fault characteristic signals, and in one embodiment, this method can subtract
Little tradition VMD decomposition algorithm updates the shadow of step-length tau vulnerable to penalty alpha, component number K and Lagrange multiplier
It rings, and the signal-to-noise ratio of output signal can be significantly increased, effectively extract Weak fault signal characteristic.
Fig. 6 and Fig. 7 are please referred to, bearing fault monitoring system module schematic diagram and bearing fault detection of the invention are shown as
System data flows to schematic diagram, as shown in Figure 6 and Figure 7, a kind of bearing fault detection system, comprising: resolution parameter module 1 is divided
Solve module 2, coupling input module 3 and stable state resonance module 4, resolution parameter module 1, to acquire original vibration signal, with grain
Swarm optimization calculates original vibration signal and obtains resolution parameter；Decomposing module 2, to Optimal Decomposition parameter, to obtain parameter combination
Information, decomposing original vibration signal accordingly is Mode Decomposition information, and decomposing module 2 is connect with resolution parameter module 1；Coupling input
Module 3, to extract the coupling input signal in Mode Decomposition information, coupling input module 3 is connect with decomposing module 2；Stable state
Resonance module 4 accordingly carries out at random altogether coupling input signal for optimizing bistable-state random resonance system with particle swarm algorithm
Vibration processing, to obtain fault characteristic signals, stable state resonance module 4 is connect with coupling input module 3.
Please refer to Fig. 8 to Figure 10, be shown as the specific module diagram in one embodiment of resolution parameter module in Fig. 6,
Original signal time domain waveform and spectrogram and particle optimizing envelope Entropy Changes schematic diagram, as shown in Fig. 8 to Figure 10, resolution parameter mould
Block 1, comprising: original signal module 11, fitness module 12, iteration setting module 13 and parameter calculating module 14, original signal
Module 11, for obtaining the original vibration signal of bearing with vibration inductor, in one embodiment, experimental verification data are used
Aero-engine Bearing testing machine carries out experiment institute to the rolling bearing with inner ring outer rollaway nest laser ablation failure geometrical characteristic
The data of acquisition.The model NU1010EM of bearing, rolling element number are 19, and added radial load is 5KN, and motor speed is
2000rpm, sample frequency 25.6KHz, the time domain waveform of the fault vibration signal of experiment acquisition and spectrogram such as Fig. 9 institute
Show, fault-signal feature is unobvious at failure-frequency in spectrogram, is not enough to support fault-signal feature extraction and determine；It adapts to
Module 12 is spent, to obtain particle iterative data according to original vibration signal, calculates fitness data, fitness module 12 accordingly
It is connect with original signal module 11；Iteration setting module 13, to according to fitness data setting population iterative logical, one
In embodiment, original vibration signal uses particle swarm algorithm Optimizing Search VMD parameter, and the search range that alpha is arranged is 100
The search range of~2500, K are the search range 0~1 of 2~10, tau, maximum number of iterations MAXIter=10, population number
N=50 is measured, iteration setting module 13 is connect with fitness module 12；Parameter calculating module 14, to be patrolled according to population iteration
It collects and calculates to obtain resolution parameter, parameter calculating module 14 is connect with iteration setting module 13.
Figure 11 and Figure 12 are please referred to, the specific module diagram in one embodiment of decomposing module in Fig. 6 and this are shown as
Levy mode function time domain waveform, as is illustrated by figs. 11 and 12, decomposing module 2, comprising: parameter component module 21 decomposes logic
Module 22 and component obtain module 23, and parameter component module 21 calculates resolution parameter, accordingly to obtain envelope data to obtain
To parameter combination information；Logic module 22 is decomposed, to decompose logic according to each parameter combination information setting mode decomposition logic
Module 22 is connect with parameter component module 21；Component obtains module 23, to according to mode decomposition logic by original vibration signal
Multiple modal components are decomposed into, in one embodiment, its optimal parameter group is obtained and is combined into (357,10,0.0672), and substitute into
VMD decomposition is carried out to original vibration signal in VMD algorithm, obtains the time domain waveform of 10 intrinsic mode function IMF, component obtains
Module 23 is connect with logic module 22 is decomposed.
Figure 13 is please referred to, the specific module diagram of coupling input module in one embodiment, such as Figure 13 in Fig. 6 are shown as
It is shown, coupling input module 3, comprising: kurtosis module 31, component screening module 32 and component processing module 33, kurtosis module 31,
To the kurtosis data of obtaining mode decomposed information, in one embodiment, the maximum IMF1 signal of kurtosis value is as coupled system
Signal to be entered；Component screening module 32, to obtain available modal components according to kurtosis data screening Mode Decomposition information, point
Amount screening module 32 is connect with kurtosis module 31；Component processing module 33 can be coupling input letter with modal components to handle
Number, component processing module 33 is connect with component screening module 32.
Please refer to Figure 14 to Figure 16, be shown as the specific module diagram in one embodiment of stable state resonance module in Fig. 6,
Controlled system output signal-to-noise ratio value variation diagram and feature enhance time domain plethysmographic signal schematic diagram, as shown in Figure 14 to Figure 16, stable state
Resonance module 4, comprising: SNR module 41 updates data module 42, resonance update module 43 and feature acquisition module 44, letter
It makes an uproar than module 41, to extract the coupled system parameter of each particle in population, calculates output signal-to-noise ratio accordingly；More new data mould
Block 42 calculates resonance system more new data accordingly, implements one speed and the position of particle to be arranged according to output signal-to-noise ratio
In example, mutative scale parameter k=3000 is taken, when particle swarm algorithm Optimal Parameters, the search range that a is arranged is the search of 0~10, b
Range is that the search range of 0~10, r is -10~10, maximum number of iterations MAXIter=100, population quantity N=50, more
New data module 42 is connect with SNR module 41；Resonate update module 43, double to be updated according to resonance system more new data
Stable state stochastic resonance system, in one embodiment, in particle swarm algorithm searching process, in per generation, corresponding controlled system exported noise
The variation of ratio is as shown in figure 15, when Figure 15 handles original vibration signal to directly adopt adaptive coupling bi-stable stochastic resonance theory,
The variation diagram of per generation corresponding controlled system output signal-to-noise ratio in optimization process, comparison it is found that the present invention to bearing Weak fault
Vibration signal has better feature reinforcing effect, and resonance update module 43 is connect with data module 42 is updated；Feature obtains mould
Block 44 handles coupling input signal acquisition fault characteristic signals to accidental resonance, in one embodiment, chooses controlled system
The corresponding parameter combination of output signal-to-noise ratio maximum value, i.e. best parameter group a=6.1824, b=3.8735, r=-1.5252；
Dual input coupling bi-stable stochastic resonance theory processing is carried out to the signal of input coupling system, obtains the time domain wave of signal after feature enhancing
Shape is as shown in figure 16, carries out spectrum analysis to Figure 16, obtains spectrogram as shown, in one embodiment, in the figure can be obvious
Find out that the spectral line at inner ring outer rollaway nest failure-frequency accounts for prominent position in entire spectrogram, almost without noise to feature extraction
It interferes, shows that the energy of non-characteristic signal is preferably converted the energy for characteristic signal by the present invention, so that feature is believed
It number is significantly increased, feature obtains module 44 and connect with the update module 43 that resonates.
In conclusion a kind of Bearing Fault Detection Method provided by the invention and system are utilized for varying input signal
Particle swarm optimization algorithm adaptively realizes the Optimum Matching of VMD algorithm parameter, so that VMD decomposition is more stable, filters out
More preferably component, avoids the subjective impact of artificial Selecting All Parameters, the present invention by two single bistable system linear couplings at
For potential well system more than one, and by the way of dual input, by coupling, realize control system and controlled system it is double altogether
Vibration is searched using system parameter and the coefficient of coup of the particle swarm algorithm to coupling bi-stable stochastic resonance theory, is significantly enhanced defeated
The signal-to-noise ratio of signal out, so that feature extraction effect is more obvious, the present invention decomposes vibration signal using VMD algorithm,
EMD and LMD method is effectively prevented to handle existing modal overlap when signal, cross envelope, owe the disadvantages of envelope, boundary effect,
The present invention by VMD algorithm with coupling bi-stable stochastic resonance theory combine, using VMD screening comprising feature-rich information component and
The advantages of coupling bi-stable stochastic resonance theory signal enhancing, can extraction Weak fault signal characteristic earlier, can be applied to other necks
Domain, with good application prospect, it is unobvious to solve resonance effect of the existing technology, parameter rely on people choose altogether and
The technical problem of signal enhancing effect difference has very high commercial value and practicability.
Claims (10)
1. a kind of Bearing Fault Detection Method characterized by comprising
Original vibration signal is acquired, the original vibration signal is calculated with particle swarm algorithm and obtains resolution parameter；
Optimize the resolution parameter, to obtain parameter combination information, decomposes the original vibration signal accordingly as Mode Decomposition letter
Breath；
Extract the coupling input signal in the Mode Decomposition information；
Optimize bistable-state random resonance system with the particle swarm algorithm, accidental resonance is carried out to the coupling input signal accordingly
Processing, to obtain fault characteristic signals.
2. the method according to claim 1, wherein it is described calculate resolution parameter the step of, comprising:
The original vibration signal of bearing is obtained with vibration inductor；
Particle iterative data is obtained according to the original vibration signal, calculates fitness data accordingly；
According to the fitness data setting population iterative logical；
The resolution parameter is calculated to obtain according to the population iterative logical.
3. the method according to claim 1, wherein the step of decomposition original vibration signal, comprising:
Envelope data is obtained, calculates the resolution parameter, accordingly to obtain the parameter combination information；
According to each parameter combination information setting mode decomposition logic；
The original vibration signal is decomposed into multiple modal components according to the mode decomposition logic.
4. the method according to claim 1, wherein the step of extraction coupling input signal, comprising:
Obtain the kurtosis data of the Mode Decomposition information；
Available modal components are obtained according to Mode Decomposition information described in the kurtosis data screening；
It can be the coupling input signal with modal components described in processing.
5. the method according to claim 1, wherein the resonance handles the step of signal, comprising:
The coupled system parameter for extracting each particle in population, calculates output signal-to-noise ratio accordingly；
Speed and the position of the particle are set according to the output signal-to-noise ratio, calculate resonance system more new data accordingly；
The bistable-state random resonance system is updated according to the resonance system more new data；
Fault characteristic signals described in the accidental resonance processing coupling input signal acquisition.
6. a kind of bearing fault detection system characterized by comprising
Resolution parameter module, to acquire original vibration signal, calculating the original vibration signal with particle swarm algorithm must be decomposed
Parameter；
Decomposing module, to obtain parameter combination information, decomposes the original vibration signal to optimize the resolution parameter accordingly
For Mode Decomposition information；
Coupling input module, to extract the coupling input signal in the Mode Decomposition information；
Stable state resonance module, it is defeated to the coupling accordingly for optimizing bistable-state random resonance system with the particle swarm algorithm
Enter signal and carry out accidental resonance processing, to obtain fault characteristic signals.
7. system according to claim 6, which is characterized in that the resolution parameter module, comprising:
Original signal module, for obtaining the original vibration signal of bearing with vibration inductor；
Fitness module calculates fitness data to obtain particle iterative data according to the original vibration signal accordingly；
Iteration setting module, to according to the fitness data setting population iterative logical；
Parameter calculating module, to calculate to obtain the resolution parameter according to the population iterative logical.
8. system according to claim 6, which is characterized in that the decomposing module, comprising:
Parameter component module calculates the resolution parameter to obtain envelope data accordingly, to obtain the parameter combination letter
Breath；
Logic module is decomposed, to according to each parameter combination information setting mode decomposition logic；
Component obtains module, the original vibration signal is decomposed into multiple mode point according to the mode decomposition logic
Amount.
9. system according to claim 6, which is characterized in that the coupling input module, comprising:
Kurtosis module, to obtain the kurtosis data of the Mode Decomposition information；
Component screening module obtains available modal components to the Mode Decomposition information according to the kurtosis data screening；
Component processing module can be the coupling input signal with modal components to handle described.
10. system according to claim 6, which is characterized in that the stable state resonance module, comprising:
SNR module calculates output signal-to-noise ratio to extract the coupled system parameter of each particle in population accordingly；
Data module is updated, speed and the position of the particle to be arranged according to the output signal-to-noise ratio, calculates resonance accordingly
System update data；
Resonate update module, to update the bistable-state random resonance system according to the resonance system more new data；
Feature obtains module, to fault characteristic signals described in the accidental resonance processing coupling input signal acquisition.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910202408.XA CN109827776B (en) | 2019-03-15 | 2019-03-15 | Bearing fault detection method and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910202408.XA CN109827776B (en) | 2019-03-15 | 2019-03-15 | Bearing fault detection method and system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109827776A true CN109827776A (en) | 2019-05-31 |
| CN109827776B CN109827776B (en) | 2024-02-13 |
Family
ID=66870277
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910202408.XA Active CN109827776B (en) | 2019-03-15 | 2019-03-15 | Bearing fault detection method and system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109827776B (en) |
Cited By (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110855374A (en) * | 2019-10-31 | 2020-02-28 | 西安交通大学 | Underwater sound target radiation noise modulation feature extraction method |
| CN111238808A (en) * | 2020-02-04 | 2020-06-05 | 沈阳理工大学 | Compound fault diagnosis method for gearbox based on empirical mode decomposition and improved variational mode decomposition |
| CN112949524A (en) * | 2021-03-12 | 2021-06-11 | 中国民用航空飞行学院 | Engine fault detection method based on empirical mode decomposition and multi-core learning |
| CN113689038A (en) * | 2021-08-24 | 2021-11-23 | 西北工业大学 | Engine cylinder fault prediction method based on multi-sensor fuzzy evaluation |
| CN114001962A (en) * | 2021-11-08 | 2022-02-01 | 重庆大学 | Method for detecting early failure of bearing by virtue of duffing stochastic resonance based on MSPSO (modeling, simulation and optimization) optimization |
| NL2028323B1 (en) * | 2021-02-03 | 2022-04-05 | Sichuan Univ Of Science And Engineering | Method for detecting internal defects of magnetic tile based on improved variational mode decomposition |
| CN115856426A (en) * | 2022-11-23 | 2023-03-28 | 吉林大学 | Particle swarm adaptive normalization-based stochastic resonance weak signal detection method |
| CN116223043A (en) * | 2023-03-29 | 2023-06-06 | 哈尔滨理工大学 | Rolling bearing weak signal detection method based on VMD and cascade stochastic resonance combination |
| CN117091648A (en) * | 2023-07-27 | 2023-11-21 | 石家庄铁道大学 | Air-ground integrated construction ecological environment monitoring device and visual processing method |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| KR20170093613A (en) * | 2016-02-05 | 2017-08-16 | 울산대학교 산학협력단 | Method for bearing fault diagnosis |
| CN107084854A (en) * | 2017-04-17 | 2017-08-22 | 四川大学 | Self-adapting random resonant Incipient Fault Diagnosis method based on grey wolf optimized algorithm |
| CN108426715A (en) * | 2018-06-13 | 2018-08-21 | 福州大学 | Rolling bearing Weak fault diagnostic method based on PSO-VMD-MCKD |
| CN108760316A (en) * | 2018-08-16 | 2018-11-06 | 苏州大学 | Information fusion method is joined in the change of variation mode decomposition |
| CN109145727A (en) * | 2018-07-11 | 2019-01-04 | 上海电力学院 | A kind of bearing fault characteristics extracting method based on VMD parameter optimization |
| CN109238715A (en) * | 2018-10-31 | 2019-01-18 | 合肥工业大学 | Bearing fault signal enhancing method and system |
| CN109257127A (en) * | 2018-09-06 | 2019-01-22 | 西安电子科技大学 | A kind of signal of communication detection method based on self-adapting random resonant |
-
2019
- 2019-03-15 CN CN201910202408.XA patent/CN109827776B/en active Active
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| KR20170093613A (en) * | 2016-02-05 | 2017-08-16 | 울산대학교 산학협력단 | Method for bearing fault diagnosis |
| CN107084854A (en) * | 2017-04-17 | 2017-08-22 | 四川大学 | Self-adapting random resonant Incipient Fault Diagnosis method based on grey wolf optimized algorithm |
| CN108426715A (en) * | 2018-06-13 | 2018-08-21 | 福州大学 | Rolling bearing Weak fault diagnostic method based on PSO-VMD-MCKD |
| CN109145727A (en) * | 2018-07-11 | 2019-01-04 | 上海电力学院 | A kind of bearing fault characteristics extracting method based on VMD parameter optimization |
| CN108760316A (en) * | 2018-08-16 | 2018-11-06 | 苏州大学 | Information fusion method is joined in the change of variation mode decomposition |
| CN109257127A (en) * | 2018-09-06 | 2019-01-22 | 西安电子科技大学 | A kind of signal of communication detection method based on self-adapting random resonant |
| CN109238715A (en) * | 2018-10-31 | 2019-01-18 | 合肥工业大学 | Bearing fault signal enhancing method and system |
Non-Patent Citations (1)
| Title |
|---|
| 王志霞等: "基于VMD的自适应随机共振在滚动轴承早期故障检测中的应用", 《机械传动》, vol. 42, no. 4, pages 144 - 149 * |
Cited By (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110855374B (en) * | 2019-10-31 | 2021-07-13 | 西安交通大学 | Underwater sound target radiation noise modulation feature extraction method |
| CN110855374A (en) * | 2019-10-31 | 2020-02-28 | 西安交通大学 | Underwater sound target radiation noise modulation feature extraction method |
| CN111238808A (en) * | 2020-02-04 | 2020-06-05 | 沈阳理工大学 | Compound fault diagnosis method for gearbox based on empirical mode decomposition and improved variational mode decomposition |
| CN111238808B (en) * | 2020-02-04 | 2021-08-17 | 沈阳理工大学 | Compound fault diagnosis method for gearbox based on empirical mode decomposition and improved variational mode decomposition |
| NL2028323B1 (en) * | 2021-02-03 | 2022-04-05 | Sichuan Univ Of Science And Engineering | Method for detecting internal defects of magnetic tile based on improved variational mode decomposition |
| CN112949524A (en) * | 2021-03-12 | 2021-06-11 | 中国民用航空飞行学院 | Engine fault detection method based on empirical mode decomposition and multi-core learning |
| CN113689038B (en) * | 2021-08-24 | 2023-05-26 | 西北工业大学 | Engine cylinder fault prediction method based on multi-sensor fuzzy evaluation |
| CN113689038A (en) * | 2021-08-24 | 2021-11-23 | 西北工业大学 | Engine cylinder fault prediction method based on multi-sensor fuzzy evaluation |
| CN114001962A (en) * | 2021-11-08 | 2022-02-01 | 重庆大学 | Method for detecting early failure of bearing by virtue of duffing stochastic resonance based on MSPSO (modeling, simulation and optimization) optimization |
| CN114001962B (en) * | 2021-11-08 | 2023-12-08 | 重庆大学 | Early failure detection method for bearing based on MSPSO optimization dufin stochastic resonance |
| CN115856426A (en) * | 2022-11-23 | 2023-03-28 | 吉林大学 | Particle swarm adaptive normalization-based stochastic resonance weak signal detection method |
| CN116223043A (en) * | 2023-03-29 | 2023-06-06 | 哈尔滨理工大学 | Rolling bearing weak signal detection method based on VMD and cascade stochastic resonance combination |
| CN117091648A (en) * | 2023-07-27 | 2023-11-21 | 石家庄铁道大学 | Air-ground integrated construction ecological environment monitoring device and visual processing method |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109827776B (en) | 2024-02-13 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109827776A (en) | Bearing Fault Detection Method and system | |
| CN110502991B (en) | Internal combustion engine health monitoring method and system based on random convolutional neural network structure | |
| CN113834656B (en) | Bearing fault diagnosis method, system, equipment and terminal | |
| Upendar et al. | Statistical decision-tree based fault classification scheme for protection of power transmission lines | |
| CN109740627A (en) | A kind of insect image identification identifying system and its method based on parallel-convolution neural network | |
| CN106500735B (en) | A kind of compressed sensing based FBG signal adaptive restorative procedure | |
| CN111523509B (en) | Equipment fault diagnosis and health monitoring method integrating physical and depth expression characteristics | |
| CN108426715A (en) | Rolling bearing Weak fault diagnostic method based on PSO-VMD-MCKD | |
| CN107884190A (en) | The Method for Bearing Fault Diagnosis decomposed based on variation mode decomposition and wavelet singular | |
| CN109711401A (en) | A kind of Method for text detection in natural scene image based on Faster Rcnn | |
| Qian et al. | An intelligent fault diagnosis method for rolling bearings based on feature transfer with improved DenseNet and joint distribution adaptation | |
| Chen et al. | Anomaly detection for drinking water quality via deep biLSTM ensemble | |
| Devulapalli et al. | Synthesized pansharpening using curvelet transform and adaptive neuro-fuzzy inference system | |
| Lai et al. | A novel nonlinear neural network ensemble model for financial time series forecasting | |
| CN111598822A (en) | Image fusion method based on GFRW and ISCM | |
| Bini et al. | An autoencoder neural network integrated into gravitational-wave burst searches to improve the rejection of noise transients | |
| Ding et al. | Mine microseismic time series data integrated classification based on improved wavelet decomposition and ELM | |
| CN109238715A (en) | Bearing fault signal enhancing method and system | |
| CN116471154A (en) | Modulation signal identification method based on multi-domain mixed attention | |
| Safavian et al. | Aquantitative comparison of different mother wavelets for characterizing transients in power systems | |
| Zheng et al. | A noise-eliminated gradient boosting model for short-term traffic flow forecasting | |
| Zhou et al. | A hybrid learning model based on auto-encoders | |
| Jane et al. | Daron: A technique for detection and removal of noise in IoT data by using central tendency | |
| CN115374687A (en) | Numerical-shape combined intelligent diagnosis method for working conditions of oil well | |
| Xiaoyun et al. | Multiple fault diagnosis for rolling bearings method employing CEEMD-GCN based on horizontal visibility graph |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |