CN103308310B - Method for demodulating impact signals on basis of all-digital peak detection - Google Patents
Method for demodulating impact signals on basis of all-digital peak detection Download PDFInfo
- Publication number
- CN103308310B CN103308310B CN201310221247.1A CN201310221247A CN103308310B CN 103308310 B CN103308310 B CN 103308310B CN 201310221247 A CN201310221247 A CN 201310221247A CN 103308310 B CN103308310 B CN 103308310B
- Authority
- CN
- China
- Prior art keywords
- value
- impact
- peak
- signal
- peak value
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Landscapes
- Testing Of Devices, Machine Parts, Or Other Structures Thereof (AREA)
- Testing Relating To Insulation (AREA)
Abstract
The invention discloses a method for demodulating impact signals on the basis of all-digital peak detection, and belongs to the field of fault diagnosis for rotary machinery. The method has the advantages that in order to overcome shortcomings of poor instantaneity of an envelope solving procedure implemented in the traditional Hilbert demodulation method and low fault detection rate under a low signal-to-noise ratio condition, an all-digital peak detection algorithm with a structural element of A*e(-t/tau) is adopted in the method and is excellent in instantaneity, applicable to embedded equipment and high in noise immunity, excellent diagnosis results can be realized when the method is applied to rotary equipment working in complicated environments of industrial fields, and the method has an important economic significance and high engineering application value.
Description
Technical field
The invention belongs to the technique study in mechanical fault diagnosis field, it is related to a kind of digital peak detection algorithm and base
Rotary machinery fault diagnosis technology in this algorithm.
Background technology
With the fast development of modern industry, various rotating machineries are widely used in each industrial circle.Rotating machinery fault
It is the major issue being related to national economy production security, early diagnosiss are carried out to fault, be related to that can equipment safe, high
The longtime running of effect, is also the normal important leverage producing of whole enterprise.Rolling bearing is quick-wear part, in many rotating machineries
Fault all relevant with rolling bearing, the fault 70% of plant equipment is vibration fault, and there are about 30% in vibration fault is by axle
Hold and cause.In addition, rolling bearing life discreteness is very big, if not carrying out fault diagnosis, and carry out more only according to projected life
Change or keep in repair, on the one hand can cause the waste of manpower financial capacity, on the other hand not up to projected life has been broken down
Bearing, it is impossible to on-call maintenance is changed, becomes potential safety hazard.Therefore monitoring, diagnosing is carried out to rolling bearing, to guarantee machine security
Run and be significant with cost-effective.Using vibration signal to diagnosing malfunction, be in rolling bearing fault diagnosis
Effectively, most common method.
Rolling bearing fault diagnosis both domestic and external mainly adopt resonance demodulation technique at present, and its principle is：When bearing is a certain
When local damage occurring on element surface, matched element surface in running, to be clashed into, it will produce impulsive force.By
Very wide in the frequency band of these shock pulse power, necessarily cover the natural frequency of monitoring component, thus the height of bearing arrangement can be evoked
Frequency intrinsic vibration, this high frequency intrinsic vibration is by the amplitude modulation(PAM) by bearing element fault characteristic frequency.By carrying out envelope
Detection, removes the frequency content of high frequency attenuation vibration, is only comprised the low frequency envelope signal of fault characteristic information, and then to this
Envelope signal is analyzed（Can be spectrum analyses it is also possible to be analyzed to peak value in the time domain）, just can recognize that the axis of rolling
The fault characteristic information holding, finally may determine that bearing fault degree or fault occur position.
The key of resonance demodulation technique is the extraction of envelope signal, and existing method mainly has：Software envelope demodulation
（Hilbert demodulation method）With hardware peak detection（gSE）.
The process of Hilbert demodulation：By Fourier transformation operation twice is carried out to primary signal, obtain envelope signal
Imaginary part, using primary signal as real part, evolution again after imaginary part and real part respectively square summation obtains demodulated signal.To demodulation
Signal carries out FFT computing, by finding characteristic frequency and its frequency multiplication of roller bearing component in frequency spectrum, thus it is fixed to realize fault
Position.The advantage of the method is to contain more complete fault message in the envelope signal obtaining.But there is also deficiency：1）Envelope
Ask for process poor real.To the only data of 8K it is necessary to carry out about 3.4e+7 time complex multiplication, so data processing complexity is very
Height, poor real, for the bearing diagnosis of low frequency heavy duty, because the data length of collection is very big, this method is less applicable.2）
In the case of low signal-to-noise ratio, fault recall rate is not high.Can obtain comprising the envelope of more complete information due to after Hilbert demodulation, and real
In the application of border, because low-frequency disturbance source is many, frequency distribution wide, the envelope signal of information completely also necessarily contains more interference
Information, especially in Incipient Fault Diagnosis, the amplitude energy of these interference signals has often flooded fault-signal, at this moment in frequency spectrum
In be difficult to identification fault characteristic frequency.
Hardware peak detection method is by designing a hardware peak detector, obtaining the zigzag envelope of primary signal
Signal, then FFT is done to this envelope signal so that it may fault location is realized by peak envelope spectrum.The principle of its peak detection
For：Vs is amplitude-modulated signal, and Vo is envelope detection signal, R>>RDTo ensure that i fills>>I is put, and that is, timeconstantτ fills<<τ is put.Work as Vs
When voltage reaches diode current flow threshold value, diode current flow charges to C, and τ fills=RDC, due to RDVery little, so τ fills very little, Vo ≈
Vs；When Vs voltage is less than diode current flow threshold value, diode ends, and C discharges through R, and τ puts=RC, because R is very big, so τ is put
Very big, the upper voltage of C declines seldom, still has Vo ≈ Vs, charge and discharge process moves in circles, and C is upper just to be obtained and envelope（Modulated signal）
Consistent voltage waveform, has the fluctuating of very little.Timeconstantτ=RC, is only determined by circuit parameter.τ is bigger, decay slower,
The transient process of circuit is longer, this is because when voltage one timing, electric capacity C is bigger, and the electric field energy of storage is bigger, is released
Discharge the required time longer；The release of electric field energy is realized by electric current again, and resistance R is bigger, and discharge current is got over
Little, discharge time is longer.In gSE method, the advantage of peak detection is：Can intactly stick signal peak value of pulse, and gSE
Judge that degree of injury is carried out in the time domain, the mathematical operation of complexity need not be carried out.But in place of its Shortcomings：Detection process
It is to utilize hardware to realize completely, wherein there is the timeconstantτ improving signal to noise ratio function, be by the resistance in circuit and capacitance
To determine so that τ-value can only select in set several gears, and to be affected by resistance and capacity cell error, its guarantor
The peak energy held also is limited by diode current flow threshold value in charging circuit.
Content of the invention
It is an object of the invention to provide a kind of impact signal demodulation method based on digital peak detection.
For reaching above-mentioned purpose, present invention employs technical scheme below.
A kind of impact signal demodulation method based on digital peak detection it is characterised in that：This impact signal demodulation side
Method simulates RC charge and discharge process, to primary signal using with A*e（-t/τ）Morphological Filtering Algorithm for structural element is filtered, when
When crude shock peakedness ratio pad value is high, crude shock peak value still as demodulation after peak value, when crude shock peakedness ratio declines
When depreciation is low or crude shock peak value is equal to pad value, abandon this peak value, continue to obtain declining of demodulated signal according to original attenuation
Subtract area, A is the impact peak value that a certain moment is more than current attenuation value, t represents the width of structural element, τ express time constant, when
Between constant, τ determine the decay speed of structural element form and peak value, τ has the effect improving signal to noise ratio.
The comprising the following steps that of described impact signal demodulation method：
Make initial impact peak A0According to A0*e(-t/τ)Decayed, obtained pad value Ai, impact that the next one is collected
Peak A1With AiRelatively, if A1Less than or equal to Ai, then abandon A1, and with AiA bit, and continue as in waveform after peak detection
Decay；If A1More than Ai, then stop this decay, by A1As the new impact peak value in waveform after peak detection, and according to
A1*e（-t/τ）Restart attenuation process.
The span of described t is 0.5T～0.8T, and T is the inaction interval of a certain part of rolling bearing.
The selection gist of described τ is：
Wherein, α represents structural element spread factor, and the span of α is 0.5～0.8, fxRepresent a certain portion of rolling bearing
The fault characteristic frequency of part, β represents remaining percentage ratio after decay, and the span of β is 10～15%.
Beneficial effects of the present invention are embodied in：
Ask for process poor real due to what traditional Hilbert demodulation method had an envelope, and low signal-to-noise ratio situation
The not high shortcoming of lower fault recall rate；And the selection of hardware peak detection existence time constant, τ is limited by circuit, and keep
Peak energy be subject to diode current flow threshold restriction shortcoming.The present invention proposes with A*e（-t/τ）Digital for structural element
Peak detection algorithm, this algorithm has good real-time, is suitably applied embedded device, has stronger anti-noise acoustic energy
Power, timeconstantτ accurately can be chosen according to fault characteristic frequency, simultaneously, it is possible to obtain more real peak energy, to industry
Slewing under live complex environment has good diagnosis effect, has important economic implications and extensive engineer applied
It is worth.
Brief description
Fig. 1 is the theory diagram of resonance and demodulation method；
Fig. 2 is hardware envelope detection schematic diagram；
Fig. 3 is structural element width to the impact figure extracting pulse number；
Fig. 4 is the process schematic of digital peak detection of the present invention；
Fig. 5 is the algorithm block diagram of digital peak detection of the present invention；
Fig. 6 (a) is bearing outer ring fault-signal, and Fig. 6 (b) is the signal after digital peak detection, and Fig. 6 (c) is demodulation
The frequency spectrum of signal.
Specific embodiment
The present invention will be further described below in conjunction with the accompanying drawings.
Referring to Fig. 1, the principle of existing resonance and demodulation method is：Original paper damages and causes impact, and impact signal evokes further
High frequency natural frequency, forms resonance, fault-signal is modulated on resonance signal, carries out envelope detection to high-frequency resonance signal band
Low frequency fault-signal can be obtained, corresponding fault diagnosis can be carried out by spectrum analyses further.
Referring to Fig. 2, existing hardware envelope detection principle is：Vs is amplitude-modulated signal, and Vo is envelope detection signal, R>>RD
To ensure that i fills>>I is put, and that is, timeconstantτ fills<<τ is put.When Vs voltage reaches diode current flow threshold value VD, diode current flow pair
C charges, and τ fills=RDC, due to RDVery little, so τ fills very little, Vo ≈ Vs；When Vs voltage is less than diode current flow threshold value, two poles
Pipe ends, and C discharges through R, and τ puts=RC, and because R is very big, so τ is put very greatly, the upper voltage of C declines seldom, still has Vo ≈ Vs, charge and discharge
Electric process moves in circles, and C is upper just to be obtained and envelope（Modulated signal）Consistent voltage waveform, has the fluctuating of very little.
Impact signal demodulation method of the present invention is based on digital peak detection, is specifically described as follows：
1. digital peak detection algorithm
Digital peak detection algorithm is based on hardware envelope detection principle, realizes detection by software algorithm, and protects
Stay the impact peak value in detection process.In fact envelope waveform depends on the voltage U at C two endsc, Uc=-iR, wherein i are that i fills or i
Put,So obtaining the differential equationSolve this single order constant coefficient linearity homogeneous differential equation, just
Obtain C both end voltage Uc=Ae（-t/RC）, wherein A is by UcInitial value determine.RC charge and discharge process is simulated by software algorithm, its
Process is to make initial impact peak A0According to A0*e（-t/τ）Decayed, obtained pad value Ai, impact value that the next one is collected
With AiRelatively, if less than equal to Ai, then abandon this peak value, and with AiA bit, and continue to decay as in waveform after detection；
If the impact peak value collecting is more than Ai, then stop this decay, using in this impact peak value waveform as after peak detection
New impact peak value, and the attenuation process again beginning the above, which achieves digital peak detection.It can be seen that it is digital
The conduction threshold of diode is not recycled to decide whether discharge and recharge in peak detection, but by comparing pad value and collecting
Subsequent time peak value size determining, referring to Fig. 4 and Fig. 5.
Digital peak detection has its advantage with respect to hardware peak detection：Ratio analog circuit high precision first, and do not have
The restriction to peak energy for the diode current flow threshold value, more can obtain real peak energy；Secondly its τ-value can arbitrarily be chosen, and
Determined by RC value not necessarily like τ-value in hardware detecting circuit, after the completion of circuit, the selection of τ-value is just certain.
See from Fig. 4, digital peak detection algorithm utilizes A*e to primary signal(-t/τ)Filtered.Therefore total
The essence of word peak detection algorithm, is to use A*e(-t/τ)The Morphological Filtering Algorithm of this structural element, wherein A be one dynamic
Value, it is the impact peak value that a certain moment is more than current attenuation value, and needs are first when carrying out shape filtering for the τ in structural element
Set, it is the key parameter determining structural element form.
2. structural element A*e(-t/τ)The selection formula of middle τ-value
Because the present invention proposes a kind of new Morphological Filtering Algorithm, so the basis of explanation shape filtering first is known here
Know.
1）Morphological operator
Shape filtering includes morphological transformation and two key elements of structural element.The basic thought of morphological transformation is to utilize structural elements
Plain this " probe ", to collect the information of signal, is constantly moved in the signal by this " probe ", signal is mated, to reach
To the purpose extracting signal, holding details and suppression noise.The mode of movement is determined by morphological operator, morphological operator is mainly wrapped
Include：Expansion, burn into open and close operator.If set A is studied object, set B is structural element, appoints in a, b difference corresponding A, B
Meaning point, then expand, burn into open and close operator is defined respectively as：
AοB=(AΘB)⊕B （3）
A·B=(A⊕B)ΘB （4）
In formula：
Ac represents supplement collection；
Br represents to negate and penetrates, that is, in B, each element negates；
What the present invention was studied is the many-valued morphological transformation of one-dimensional signal, and corresponding 4 kinds of morphological operator can so define,
Assume that list entries f (n) are to be defined on integer domain of definition F=0, the discrete function on 1 ..., N-1, definition structure element sequence g
N () is domain of definition G=0, the discrete function on 1 ..., M-1, and N >=M, m ∈ 0,1 ..., M-1, then f (n) is swollen with regard to g (n)
Swollen, corrosion and open and close operator be defined as：
(f⊕g)(n)=max[f(n-m)+g(m)] （5）
(fΘg)(n)=min[f(n+m)-g(m)] （6）
(fοg)(n)=(fΘg⊕g)(n) （7）
(f·g)(n)=(f⊕gΘg)(n) （8）
Dilation operation eliminates the negative pulse of signal and has smoothed positive pulse；Erosion operation eliminates positive pulse and smooths
Negative pulse；Opening operation eliminates positive pulse and remains negative pulse；Closed operation eliminates negative pulse and remains positive pulse.
2）Structural element
The effect of shape filtering depends not only on adopted morphological operations, but also have with the structural elements being adopted
Close, the size and dimension of structural element will be as close possible to the signal aspect being gathered.The selection of structural element mainly includes tying
The type of constitutive element, amplitude and width.Type can be straight line, triangle, square, circle etc..The present invention is according to the axis of rolling
Hold the Morphological Features of fault-signal, using A*e（-tτ）As structural element, its amplitude is dynamically determined by A, i.e. current impact peak value
Fixed, then the width of structural element to be determined by timeconstantτ the most at last.
The change width of structural element can lead to the pulse number extracting different.Narrower the extracted arteries and veins of structural element
Rush that number is more, the fault message comprising is just many, but the noise signal being mixed into is also many；Wider the extracted pulse of structural element
Number is fewer, and while compacting noise signal, some important fault messages are likely to lose, and is equally unfavorable for improving noise
Than.According to Fig. 3, Nikolaou N.G finds out to the analysis result of linear structure element, and the change width of structural element can lead to
The pulse number extracting is different.Narrower the extracted pulse number of structural element is more, and the fault message comprising is just many, but
The noise signal being mixed into is also many；Wider the extracted pulse number of structural element is fewer, while compacting noise signal, some weights
The fault message wanted is likely to lose, and is equally unfavorable for improving signal to noise ratio.As shown in Figure 3, structural element width is in 0.5T
～0.8T（T is the inaction interval of a certain part of bearing）Between when, the down pulse number of extraction at most, and noise pulse number
Minimum, i.e. filtered signal to noise ratio highest.
According to the analysis result to linear structure element for the Nikolaou N.G, for obtaining optimal filter effect, the present invention
Structural element A*e（-t/τ）The preferred 0.6T of width, i.e. t=0.6T.Because digital peak detection is based on hardware circuit RC charge and discharge
Electricity principle, thus in digital peak detection present percussion peak A decay it should with RC zero input response transient process class
Seemingly, in theory t=∞ when RC zero input response transient process terminate, but actually when peak atenuation 85%～90% is considered as transient state
Process terminates substantially, but because rolling bearing size is different with the working condition in industry spot, such as load, live noise is done
Disturb, A attenuation should be variant, after note A decay, remaining percentage ratio is β, then obtains formula：
β·A=A*e（-0.6T/τ）（9）
In formulafxIt is the fault characteristic frequency of a certain part of rolling bearing.
Abbreviation formula（9）Can obtain：
By formula（10）The solving equation of τ-value can be obtained, such as deriving（11）：
Under normal circumstances, β can select 15%.
It can be seen from figure 4 that demodulated signal includes peak value and decay area, when crude shock peakedness ratio pad value is high,
Using still as the peak value after demodulation, and when crude shock peakedness ratio pad value is low or crude shock peak value is equal to pad value, get rid of
Abandon this peak value, continue to obtain the decay area of demodulated signal according to original attenuation.After demodulation, original high-frequency resonance signal is filtered
Remove, retain the low frequency signal related to fault.
Referring to Fig. 5, digital peak detection algorithm makes initial impact peak A0According to A0*e(-t/τ)Decayed, declined
Depreciation Ai, the impact value that collect the next one and AiRelatively, if less than equal to Ai, then abandon this peak value, and with AiAs
In waveform after detection a bit, and continue to decay；If the impact value collecting is more than Ai, then stop this decay, this rushed
Hit value as the new impact peak value in waveform after peak detection, and the attenuation process again beginning the above, which achieves total
Word peak detection.
Referring to Fig. 6（a）, obvious impact peak value can be seen from original bearing outer ring fault-signal, these impact peak
Value is caused by outer ring fault, by digital peak detection, is only comprised the demodulated signal of low frequency fault message, referring to
Fig. 6（b）, FFT computing is carried out to demodulated signal, obtains demodulated signal spectrogram, referring to Fig. 6（c）, from frequency spectrum, can be clear
See bearing fault frequency and its frequency multiplication it was demonstrated that the effectiveness of the inventive method.Compared with Hilbert demodulation, by complete
Signal after digital peak demodulation has very high signal to noise ratio, can be it can be clearly seen that bearing fault is special in demodulated signal frequency spectrum
Levy frequency and its frequency multiplication, and algorithm complex is greatly lowered, there is more preferable real-time and capacity of resisting disturbance.With the inspection of hardware peak value
Ripple is compared, and more accurately maintains the impact peak value of primary signal, eliminates the hardware circuit of complexity meanwhile, reduces cost.
Above content is to further describe it is impossible to assert with reference to specific preferred implementation is made for the present invention
The specific embodiment of the present invention is only limitted to this, for general technical staff of the technical field of the invention, is not taking off
On the premise of present inventive concept, some simple deduction or replace can also be made, all should be considered as belonging to the present invention by institute
The claims submitted to determine scope of patent protection.
Claims (1)
1. a kind of impact signal demodulation method based on digital peak detection it is characterised in that：This impact signal demodulation method
Simulation RC charge and discharge process, to primary signal using with A*e(-t/τ)Morphological Filtering Algorithm for structural element is filtered, when former
When beginning impact peak value specific damping value is high, crude shock peak value still as demodulation after peak value, when crude shock peak value specific damping
When being worth low or crude shock peak value equal to pad value, abandon this peak value, continue to obtain the decay of demodulated signal according to original attenuation
Area；A is the impact peak value that a certain moment is more than current attenuation value, and t represents the width of structural element, τ express time constant；
The value of described t is 0.6T, and T is the inaction interval of a certain part of rolling bearing；
The selection gist of described τ is：
Wherein, α represents structural element spread factor, and the value of α is 0.6, fxRepresent the fault signature frequency of a certain part of rolling bearing
Rate, β represents remaining percentage ratio after decay, and the value of β is 15%；
The comprising the following steps that of described impact signal demodulation method：
Make initial impact peak A0According to A0*e(-t/τ)Decayed, obtained pad value Ai, impact peak value that the next one is collected
A1With AiRelatively, if A1Less than or equal to Ai, then abandon A1, and with AiA bit, and continue to decay as in waveform after peak detection；
If A1More than Ai, then stop this decay, by A1As the new impact peak value in waveform after peak detection, and according to A1*e(-t/τ)Restart attenuation process.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201310221247.1A CN103308310B (en) | 2013-06-05 | 2013-06-05 | Method for demodulating impact signals on basis of all-digital peak detection |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201310221247.1A CN103308310B (en) | 2013-06-05 | 2013-06-05 | Method for demodulating impact signals on basis of all-digital peak detection |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN103308310A CN103308310A (en) | 2013-09-18 |
| CN103308310B true CN103308310B (en) | 2017-02-08 |
Family
ID=49133775
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201310221247.1A Active CN103308310B (en) | 2013-06-05 | 2013-06-05 | Method for demodulating impact signals on basis of all-digital peak detection |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN103308310B (en) |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103940597A (en) * | 2014-04-08 | 2014-07-23 | 昆明理工大学 | Method for detecting mechanical faults based on generalized extremum morphological filtering |
| CN109063668B (en) * | 2018-08-14 | 2021-09-14 | 哈工大机器人（合肥）国际创新研究院 | Impact signal envelope demodulation method based on peak value retention and down-sampling |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102183366A (en) * | 2011-03-08 | 2011-09-14 | 上海大学 | Device and method for vibration measurement and failure analysis of rolling bearing |
| CN102426102A (en) * | 2011-10-19 | 2012-04-25 | 唐德尧 | Resonance demodulation double isolate frequency spectrum method for detecting crack of gear shaft |
Family Cites Families (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JPH05281094A (en) * | 1992-03-31 | 1993-10-29 | Nkk Corp | Inspection method of directly moving type guide provided with circulating type rolling guide and its apparatus |
| JP3264480B2 (en) * | 1996-11-28 | 2002-03-11 | 川崎製鉄株式会社 | Abnormality diagnosis method for wheels with built-in bearings and low-speed rotating bearings |
-
2013
- 2013-06-05 CN CN201310221247.1A patent/CN103308310B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102183366A (en) * | 2011-03-08 | 2011-09-14 | 上海大学 | Device and method for vibration measurement and failure analysis of rolling bearing |
| CN102426102A (en) * | 2011-10-19 | 2012-04-25 | 唐德尧 | Resonance demodulation double isolate frequency spectrum method for detecting crack of gear shaft |
Non-Patent Citations (2)
| Title |
|---|
| 振动信号处理技术及其在机械故障诊断中的应用;冯坤;《中国优秀硕士学位论文全文数据库工程科技Ⅱ辑》;20061115(第11期);正文第43-48页 * |
| 谈两种信号下二极管检波;房淑芬;《科技创新导报》;20091231(第12期);第63-64页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN103308310A (en) | 2013-09-18 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Yan et al. | Fault diagnosis of rolling element bearing using a new optimal scale morphology analysis method | |
| Smith et al. | Optimised Spectral Kurtosis for bearing diagnostics under electromagnetic interference | |
| CN103439110B (en) | Rolling bearing early-stage weak fault diagnostic method | |
| CN104677623B (en) | A kind of blade of wind-driven generator fault acoustics in place diagnostic method and monitoring system | |
| CN103197171B (en) | Based on the method for the characteristic parameter assessment oil paper insulation ageing state of time domain dielectric response | |
| CN102269655B (en) | Method for diagnosing bearing fault | |
| CN103398769A (en) | Transformer on-line fault detecting method based on sampling integrated SVM (support vector machine) under wavelet GGD (general Gaussian distribution) feather and unbalanced K-mean value | |
| CN104792528A (en) | Adaptive optimal envelope demodulation method | |
| CN103901111A (en) | Nondestructive detection system and method of blades of wind turbine generator system | |
| CN106932184A (en) | A kind of Diagnosis Method of Transformer Faults based on improvement hierarchical clustering | |
| CN104950230A (en) | Power distribution network fault line selection method based on variable-scale bi-stable system | |
| Li et al. | A VME method based on the convergent tendency of VMD and its application in multi-fault diagnosis of rolling bearings | |
| CN101886977A (en) | Method for self-adaptively detecting periodic instant component in signal | |
| CN103018632A (en) | Small current grounding system single-phase ground fault line selection method based on fisher information | |
| CN103308310B (en) | Method for demodulating impact signals on basis of all-digital peak detection | |
| CN106324406A (en) | Transformer direct-current magnetic bias fault diagnosis method and device | |
| Berntsen et al. | Enhanced demodulation band selection based on Operational Modal Analysis (OMA) for bearing diagnostics | |
| Zhang et al. | An adaptive demodulation band segmentation method to optimize spectral boundary and its application for wheelset-bearing fault detection | |
| CN104965155B (en) | A kind of wave head lookup method based on filtering characteristic | |
| CN102778354A (en) | Wind power slewing bearing fault diagnosis method based on wavelet analysis | |
| Wang et al. | Fault diagnosis of rolling bearings based on enhanced optimal morphological gradient product filtering | |
| CN209783909U (en) | Rolling bearing multi-stage fault pre-judging system | |
| CN102680080B (en) | Unsteady-state signal detection method based on improved self-adaptive morphological filtering | |
| CN205209627U (en) | Power transformer winding axial compresses tightly state on -line monitoring system | |
| CN104156412A (en) | Complex event processing based power quality disturbance event classification monitoring method |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant |