CN109655266A - A kind of Wind turbines Method for Bearing Fault Diagnosis based on AVMD and spectral coherence analysis - Google Patents
A kind of Wind turbines Method for Bearing Fault Diagnosis based on AVMD and spectral coherence analysis Download PDFInfo
- Publication number
- CN109655266A CN109655266A CN201811507395.9A CN201811507395A CN109655266A CN 109655266 A CN109655266 A CN 109655266A CN 201811507395 A CN201811507395 A CN 201811507395A CN 109655266 A CN109655266 A CN 109655266A
- Authority
- CN
- China
- Prior art keywords
- fault
- signal
- avmd
- spectrum
- component
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01M—TESTING STATIC OR DYNAMIC BALANCE OF MACHINES OR STRUCTURES; TESTING OF STRUCTURES OR APPARATUS, NOT OTHERWISE PROVIDED FOR
- G01M13/00—Testing of machine parts
- G01M13/04—Bearings
Landscapes
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Testing Of Devices, Machine Parts, Or Other Structures Thereof (AREA)
Abstract
The invention discloses a kind of Wind turbines Method for Bearing Fault Diagnosis based on AVMD and spectral coherence analysis, it is inaccurate for the discomposing effect of conventional bearings method for diagnosing faults at this stage, the low problem of computational efficiency proposes a kind of new large-scale wind electricity unit Method for Bearing Fault Diagnosis.The core concept of algorithm is: decomposing different types of fault-signal by improved VMD algorithm i.e. AVMD algorithm first, mode is decomposed using the denoising of PCA dimensionality reduction later, PCA treated principal component is subjected to Spectrum Conversion, it goes to frequency domain and by the end to end spectral vectors for obtaining fault signature of the frequency spectrum of each mode of gained, constructs knowledge base.Signal to be detected is subjected to same treatment later.Fault diagnosis is completed using spectral coherence analysis method.This method is more accurate compared to traditional Wind turbines Fault Diagnosis of Roller Bearings, and calculating speed faster, there is preferable practical value.
Description
Technical field
The present invention is a kind of fault diagnosis side applied to Wind turbines rolling bearing or heavy mechanical equipment rolling bearing
Method reinforces signal decomposition accuracy especially for vibration signal non-stationary, nonlinear characteristic；Belong to based on data-driven
Fault diagnosis technology field.
Background technique
As the mankind require to continue to increase to the energy, power industry is developed rapidly, and wind power industry is opposite with its cost
Lower, wind resource is abundant, and the advantages such as energy source green and environment-friendly have become the Main Trends of The Development of clean energy resource.And wind power plant is built
Directly decide that the benefit of wind power plant, wind-power electricity generation project are a kind of engineerings that investment time is very long at maintenance cost later,
Probably at 7 years or so, and the income phase is also very long, or even can be more than 10 years.The wind power generating set operated for a long time needs regular
It is overhauled and is safeguarded, to ensure the stability and safety that run.When the working life of wind power generating set in engineering is 20
Nian Shi, maintenance cost just account for the 10%~15% of integral benefit；Operation required for wind power generating set installation at sea
The 20%~25% of overall efficiency is accounted for maintenance cost, a large amount of operating and maintenance cost increase engineering running cost and
Reduce the economic well-being of workers and staff of engineering.Make the maximizing the benefits of wind power plant, it is necessary to be preferably minimized O&M cost.Rolling bearing
It is one of important source of trouble in one of vital component of wind-driven generator and gear case of blower.It shows according to statistics, it is mechanical
About 30% failure is all as caused by rolling bearing in failure, and electrical fault, which also has 20% failure, to be caused by rolling bearing
's.In addition most of Wind turbines are mounted in the more sufficient area of wind resource, such as grassland, Gobi desert in Practical Project
With the environment such as desert, the range of units' installation is very wide and quantity also compares more, is influenced by severe natural environment, this makes
Rolling bearing is easier to break down.Fan trouble gently then causes the loss of electric power energy once handling not in time, heavy then cause
Machinery equipment scrap and casualties.Therefore, fault diagnosis is completed to Wind turbines rolling bearing in time to have great significance.
The vibration signal near blower rolling bearing is analyzed, and then completing fault diagnosis is a kind of good method.But
Non-stationary, nonlinear characteristic is often presented in its vibration signal, so that the fault message sufficiently excavated in signal becomes difficult.It seeks
It is vital for looking for a kind of suitable signal analysis method.At present the analysis signal method that has proposed have wavelet analysis,
EMD, EEMD etc..Signal characteristic is obtained by the signal time frequency analysis of front end, has been come in conjunction with suitable rear end mode identification method
At fault diagnosis.VMD is a kind of signal Time-Frequency Analysis Method newly proposed, and this method is excellent in the decomposition to different frequency component
In wavelet analysis, EMD, EEMD method is several signal processing method spectral contrast figures as shown in fig. 1 to fig. 4.But routine VMD
Algorithm decomposes mode quantity N and penalty parameter ε and needs to be manually set, if the selection of two parameters it is bad on discomposing effect influence compared with
Greatly.It for this problem, needs to improve traditional VMD, completes failure in conjunction with the method for suitable pattern-recognition later
Diagnosis.
Summary of the invention
The present invention is inaccurate for the discomposing effect of conventional bearings method for diagnosing faults at this stage, and computational efficiency is low to ask
Topic proposes a kind of new large-scale wind electricity unit Method for Bearing Fault Diagnosis.The core concept of algorithm is: first by improved
VMD algorithm, that is, AVMD algorithm decomposes different types of fault-signal, decomposes mode using the denoising of PCA dimensionality reduction later, PCA is handled
Rear principal component carries out Spectrum Conversion, goes to frequency domain and by the end to end frequency spectrum for obtaining fault signature of the frequency spectrum of each mode of gained
Vector constructs knowledge base.Signal to be detected is subjected to same treatment later.Failure is completed using spectral coherence analysis method to examine
It is disconnected.This method is more accurate compared to traditional Wind turbines Fault Diagnosis of Roller Bearings, and calculating speed faster, there is preferable reality
With value.
Present invention employs technical solution be a kind of Wind turbines bearing fault based on AVMD and spectral coherence analysis examine
Disconnected method, the realization of this method include the following steps:
A. the fault database stage is established:
The acquisition of step 1) data and division: the original vibration signal of acquisition M class failure draws kth class fault-signal
Point, k=1,2 ..., M, every kind of fault-signal are divided into H sections, and every segment length is consistent.Two kinds of diagnosis algorithms are used as S sections before choosing
Training sample be trained, after selection K segment data as detection sample verification algorithm validity, wherein 0 < S+K≤H；
Step 2) parameter is chosen: Short Time Fourier Transform carried out to original signal first, draws Fourier spectrum figure in short-term,
Mode number N is obtained by image result.Gradually change the size of penalty parameter ε since 5 to original using amendment step-length τ=50 later
Signal is decomposed, and the superposition frequency spectrum of decomposition result and original signal spectrum is carried out mutually closing property analysis, highest relative coefficient is corresponding
Penalty parameter ε be required.
Step 3) extracts failure puppet spectrum signature: being decomposed first using AVMD to M class fault-signal, each failure letter
Number decomposition obtains N number of modal components.Using the decomposition mode of PCA processing AVMD, N number of principal component is obtained, chooses contribution rate highest
Preceding n principal component, n < N achievees the purpose that dimensionality reduction and denoising.And this n principal component is subjected to Fast Fourier Transform (FFT), it
Pseudo- spectrum signature vector is obtained by obtained frequency spectrum is end to end in order afterwards, and this puppet spectrum signature vector is for cross-correlation point
Analysis.
B. implement diagnostic phases:
Failure puppet spectrum signature vector is obtained to unknown failure signal, by the puppet spectrum signature vector respectively with fault signature
The faulty pseudo- spectrum signature vector of institute in library seeks cross-correlation coefficient, and fault type corresponding to maximum cross-correlation coefficient is to examine
Disconnected result.It is as follows that cross-correlation coefficient r seeks formula:
Wherein X represents the pseudo- spectrum signature vector of unknown failure signal, a failure classes in Y representing fault characteristic set
The pseudo- spectrum signature vector of type.Cov represents the covariance for seeking two pseudo- spectrum signature vectors.σ represents standard deviation.Failure of the same race
Characteristic frequency has very strong correlation.Relative coefficient range is -1 to 1, | r | value illustrate that correlation is bigger closer to 1,
Illustrate that correlation is smaller or even uncorrelated closer to 0.
Compared with prior art, the present invention proposes that a kind of blower rolling bearing fault diagnosis of AVMD- spectral coherence analysis is new
Method.The method overcome the defects that common VMD algorithm manually chooses mode number Yu two parameters of penalty parameter.At PCA
Reason AVMD decomposition result has reached dimensionality reduction denoising, the accurate purpose for choosing fault signature.Finally event is compared using spectral coherence analysis
Hindering characteristic spectrum model library and signal spectrum signature to be detected, the corresponding fault type of screening maximum correlation coefficient completes fault diagnosis,
Arithmetic speed is fast, improves diagnosis efficiency.
Detailed description of the invention
Fig. 1 is VMD decomposed signal spectrogram.
Fig. 2 is EMD decomposed signal spectrogram.
Fig. 3 is EEMD decomposed signal spectrogram.
Fig. 4 is wavelet decomposition signal spectrum figure.
Fig. 5 is malfunction test platform.
Fig. 6 is algorithm specific flow chart.
Fig. 7 is the Fourier spectrum figure in short-term for selecting mode number.
Fig. 8 is optimal penalty parameter selection figure.
Fig. 9 is that the real component of penalty parameter ε=100 is superimposed frequency spectrum and original signal spectral contrast figure.
Figure 10 is that the real component of penalty parameter ε=2000 is superimposed frequency spectrum and original signal spectral contrast figure.
Figure 11 is that the real component of penalty parameter ε=5000 is superimposed frequency spectrum and original signal spectral contrast figure.
Figure 12 is each principal component contributor rate figure.
Figure 13 is outer ring characteristic frequency spectrum and inner ring characteristic frequency spectrum comparison diagram after being handled using PCA.
Figure 14 is 0.007 outer ring fault diagnosis result figure.
Figure 15 is 0.007 inner ring fault diagnosis result figure.
Figure 16 is 0.007 ball fault diagnosis result figure.
Figure 17 is 0.021 outer ring fault diagnosis result figure.
Figure 18 is 0.021 inner ring fault diagnosis result figure.
Figure 19 is 0.021 ball fault diagnosis result figure.
Figure 20 is wind field data outer ring fault diagnosis result figure.
Figure 21 is wind field data inner ring fault diagnosis result figure.
Specific embodiment
The purpose of the present invention is insufficient mainly for the accuracy of conventional rolling bearing method for diagnosing faults diagnostic result, calculates
Amount is big, the low problem of diagnosis efficiency.
A kind of Wind turbines Method for Bearing Fault Diagnosis based on AVMD and spectral coherence analysis, feature include " fault database
Modeling " and " implementing diagnosis " two stages, the specific steps are as follows:
A. the fault database stage is established:
1) fault-signal catabolic phase；
Fault-signal is decomposed into N number of modal components using adaptive variation mode decomposition method (AVMD), for M type
The fault-signal of type reuses AVMD and decomposes every kind of fault-signal, is divided into solution M times.Obtain the signal component that M group number is N
Set；
2) characteristic frequency spectrum obtains the stage；
Principal component analysis (PCA) is carried out to the modal components of M class fault-signal first and denoises dimension-reduction treatment, chooses contribution rate
Highest preceding n principal component component is translated into frequency spectrum using Fast Fourier Transform (FFT) (FFT).By n principal component component
Frequency spectrum is end to end, constitutes a fault signature puppet frequency spectrum.M class fault-signal is needed to construct event to every class fault-signal respectively
Hinder feature puppet spectral vectors.Set A comprising this M feature puppet spectral vectors is known as fault signature set.
B. implement diagnostic phases:
For unknown failure signal, N number of intrinsic modal components are broken down into using AVMD method, using PCA by N number of
Sign modal components dimensionality reduction denoising is converted into N number of principal component component, using Fast Fourier Transform (FFT) by the highest preceding n master of contribution rate
Ingredient component becomes frequency spectrum, and end to end n1 frequency spectrum constitutes pseudo- spectrum signature vector.By the puppet spectrum signature vector respectively with
The faulty pseudo- spectrum signature vector of institute in fault signature set A asks cross-correlation coefficient, event corresponding to maximum cross-correlation coefficient
Hindering type is diagnostic result.It is as follows that cross-correlation coefficient r seeks formula:
Wherein X represents the pseudo- spectrum signature vector of unknown failure signal, one of event in Y representing fault characteristic set
Hinder the pseudo- spectrum signature vector of type.Cov represents the covariance for seeking two pseudo- spectrum signature vectors.σ represents standard deviation.
1. the determination of mode number.Before by signal variation mode decomposition, drawn in short-term using Short Time Fourier Transform
Fourier transformation spectrogram obtains mode number N according to image.
2. the determination of penalty parameter ε.Penalty parameter in variation mode decomposition is changed using amendment step-length τ, will be decomposed every time
Mode result afterwards merges, and seeks cross-correlation coefficient with original signal.It is available optimal according to cross-correlation coefficient point of inflexion on a curve
Penalty parameter ε.
3. decomposing fault-signal using AVMD.According to determining intrinsic mode number N and optimal penalty parameter ε, AVMD is set
Parameter decomposes fault-signal using AVMD later and obtains the component signal of different modalities.
Embodiment
The present invention successively uses the validity of two kinds of data proof algorithms.Primary sources use laboratory blower transmission chain
Platform acquires bearing outer ring, bearing inner race, bearing ball fault data and operates normally data；Secondary sources acquire wind-force hair
The bearing inner race of the true Wind turbines of electric field, operates normally data at bearing outer ring.
The following are two kinds of data related introductions:
Lab platform bearing data experiment data are to process Single Point of Faliure on bearing by spark technology, bearing class
Type is SKF6205, measures bearing vibration signal with acceleration transducer.The multi-group data of different condition is contained in data, is selected
It selects the bearing driving end vibration signal that load is 3HP, revolving speed 1730rpm, sample frequency are 12000Hz and carries out proof of algorithm.
The corresponding bearing operating status of data used includes normal, inner ring failure, outer ring failure and four seed type of rolling element failure, damage
Hurt diameter and uses 0.007 inch and 0.021 inch of two class.Fig. 5 is malfunction test platform.
Inner Mongol Hui Tengliang wind power plant (all blower models are the bright 1.5MW blower of the sun) wind-driven generator of acquisition back
Bearing fault data are divided into outer ring failure, inner ring failure and normal signal three classes data, sample frequency 26kHz, bearing designation
For the deep groove ball bearing of 6332MC3SKF.Rolling bearing design parameter is as shown in table 1.
1 rolling bearing 6332MC3SKF basic parameter of table
The method of the present invention is realized into rolling bearing fault diagnosis, main includes establishing fault database and implementing to diagnose two big step
Suddenly, it if Fig. 6 is specific flow chart of the invention, is specifically presented below:
A. the fault database stage is established:
Step 1:, will be 12000 in the original signal data of every kind of failure for every kind of fault-signal of experiment porch data
A point is divided into 300 samples, and each sample includes 400 points.For every kind of fault-signal of wind field data, by every kind of failure
12000 points are divided into 300 samples in original signal data, and each sample includes 400 points.
Step 2: the mode number in order to determine fault-signal, it will be former using Short Time Fourier Transform function (tfrstft)
Barrier signal is converted to Fourier spectrum figure in short-term, is determined by taking 0.007 inner ring fault-signal as an example according to Fourier spectrum figure in short-term
Mode number N=4 is decomposed, mode selection course is as shown in Figure 7.The frequency y direction for observing Fourier spectrum figure in short-term, can
To find out that signal contains 4 frequency component parts.
Step 3: setting penalty parameter initial value ε=5, since correcting step-length τ=50.Set cycle-index initial value ii=
1, recycle total degree nn=100.Using VMD preliminary exposition fault-signal, the component after decomposition is respectively subjected to fast Fourier
Transformation, and it is superimposed obtained component frequency spectrum.Component superposition frequency spectrum and original signal frequency spectrum are subjected to correlation analysis later, obtain one
A relative coefficient r1.Ii=ii+1 enters cyclic process next time, until ii=nn.Finally according to relative coefficient figure
Inflection point filter out the corresponding penalty factor of maximal correlation property coefficient.Relative coefficient calculation formula is as follows:
X (t) and y (t) respectively represents original signal frequency spectrum and is superimposed frequency spectrum with component in formula, and σ represents standard deviation.
By taking 0.007 inner ring fault-signal as an example, the penalty parameter selection course of AVMD as shown in figure 8, penalty parameter from 0 to
In 4000 conversion process, it can be seen that component superposition frequency spectrum and original signal as penalty factor ε=2000, after decomposition
Frequency spectrum correlation highest.Such as Fig. 9, Figure 10, shown in Figure 11, when to select penalty parameter being respectively 100,2000 and 5000, in 0.007
Enclose the comparison diagram of fault component superposition frequency spectrum and original signal frequency spectrum, it can be seen that when penalty parameter is 2000, two kinds of frequency spectrums are overlapped
Highest is spent, shows that discomposing effect is most accurate when taking penalty parameter is 2000.
Step 4: being handled using the modal components that PCA algorithm decomposes M class fault-signal, obtain 4 principal components, schemed
12 be the contribution rate of each principal component, chooses highest preceding 3 principal components of contribution rate and is carried out Fast Fourier Transform (FFT) conversion
For frequency spectrum, pseudo- spectrum signature vector is constituted according to these frequency spectrums of sequence tandem array of contribution rate from high to low.As shown in figure 13
It is the comparison diagram of inner ring puppet spectrum signature vector Yu outer ring puppet spectrum signature vector, it can be seen that wherein different feature vectors tool
There is apparent difference.M class puppet spectrum signature vector is constituted into set A, A is spectrum signature library；
B. implement the fault diagnosis stage:
The same treatment that unknown failure signal is carried out to step 1~4 obtains pseudo- spectrum signature vector to be detected, this is waited for
It detects the M class failure puppet spectrum signature vector that pseudo- spectrum signature vector is obtained with step 4 respectively and carries out cross-correlation analysis, obtain M
A cross-correlation coefficient r, wherein the corresponding fault type of maximum r value is to diagnose final result.
Above-mentioned steps are concrete application of the method for the present invention in rolling bearing fault.In order to verify the effective of this method
Property, to 0.007 outer ring in testing stand data, 0.007 inner ring, 0.007 ball, 0.021 outer ring, 0.021 inner ring, 0.021 rolling
Pearl fault data has carried out fault diagnosis experiment, and inner ring data and outer ring data in wind field data is used to carry out event again later
Hinder diagnostic test.Using experiment porch data obtain experimental result is shown in Figure 14 to Figure 19, obtained using true blower data
Diagnostic result is shown in Figure 20 to Figure 21.A possibility that every width figure respectively includes each fault type size curve, the wherein height of curve
Degree represents a possibility that being diagnosed as the type failure size, and curve is higher to illustrate a possibility that unknown signaling is the type failure more
Greatly.By the visible AVMD- correlation analysis of Figure 14 to Figure 21 to the diagnostic result curve of various fault types without intersection, diagnosis
It works well.As shown in table 2, the correlation analysis side VMD- algorithm, EMD- relevance algorithms are used respectively, and EEMD- correlation is calculated
Method, wavelet decomposition-correlation analysis algorithm carry out fault diagnosis.Wherein VMD- correlation analysis algorithm is very fast, high-efficient.Though
Right wavelet decomposition-correlation analysis algorithm diagnoses speed also quickly, but frequency of the wavelet decomposition to signal known to Fig. 1 and Fig. 4
Domain discomposing effect exists obvious insufficient.Therefore the method for the present invention has powerful advantage on discomposing effect and computational efficiency.
2 four kinds of algorithm operation time comparisons of table
Claims (2)
1. a kind of Wind turbines Method for Bearing Fault Diagnosis based on AVMD and spectral coherence analysis, it is characterised in that: this method packet
Include " fault database modeling " and " implementing diagnosis " two stages, the specific steps are as follows:
A. the fault database stage is established:
1) fault-signal catabolic phase；
Fault-signal is decomposed into N number of modal components using adaptive variation mode decomposition method AVMD, the event for M seed type
Hinder signal, reuses AVMD and decompose every kind of fault-signal, be divided into solution M times；Obtain the signal component set that M group number is N；
2) characteristic frequency spectrum obtains the stage；
Principal component analysis PCA is carried out to the modal components of M class fault-signal first and denoises dimension-reduction treatment, it is highest to choose contribution rate
Preceding n principal component component is translated into frequency spectrum using Fast Fourier Transform (FFT) FFT；By the frequency spectrum head and the tail of n principal component component
Connect, constitutes a fault signature puppet frequency spectrum；M class fault-signal is needed pseudo- to every class fault-signal building fault signature respectively
Spectral vectors；Set A comprising this M feature puppet spectral vectors is known as fault signature set；
B. implement diagnostic phases:
For unknown failure signal, N number of intrinsic modal components are broken down into using AVMD method, using PCA by N number of eigen mode
The denoising of state component dimensionality reduction is converted into N number of principal component component, using Fast Fourier Transform (FFT) by the highest preceding n principal component of contribution rate
Component becomes frequency spectrum, and end to end n1 frequency spectrum constitutes pseudo- spectrum signature vector；By the puppet spectrum signature vector respectively with failure
The faulty pseudo- spectrum signature vector of institute in characteristic set A asks cross-correlation coefficient, failure classes corresponding to maximum cross-correlation coefficient
Type is diagnostic result；It is as follows that cross-correlation coefficient r seeks formula:
Wherein X represents the pseudo- spectrum signature vector of unknown failure signal, one of failure classes in Y representing fault characteristic set
The pseudo- spectrum signature vector of type；Cov represents the covariance for seeking two pseudo- spectrum signature vectors；σ represents standard deviation.
2. a kind of Wind turbines Method for Bearing Fault Diagnosis based on AVMD and spectral coherence analysis according to claim 1,
It is characterized by: the 1. determination of mode number；Before by signal variation mode decomposition, drawn using Short Time Fourier Transform short
When Fourier transformation spectrogram, according to image obtain mode number N；
2. the determination of penalty parameter ε；Penalty parameter in variation mode decomposition is changed using amendment step-length τ, after each decompose
Mode result merges, and seeks cross-correlation coefficient with original signal；Optimal ginseng is penalized according to cross-correlation coefficient point of inflexion on a curve is available
Number ε；
3. decomposing fault-signal using AVMD；According to determining intrinsic mode number N and optimal penalty parameter ε, AVMD parameter is set,
Fault-signal, which is decomposed, using AVMD later obtains the component signal of different modalities.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811507395.9A CN109655266B (en) | 2018-12-11 | 2018-12-11 | Wind turbine generator bearing fault diagnosis method based on AVMD and spectrum correlation analysis |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811507395.9A CN109655266B (en) | 2018-12-11 | 2018-12-11 | Wind turbine generator bearing fault diagnosis method based on AVMD and spectrum correlation analysis |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109655266A true CN109655266A (en) | 2019-04-19 |
| CN109655266B CN109655266B (en) | 2021-12-10 |
Family
ID=66113254
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811507395.9A Active CN109655266B (en) | 2018-12-11 | 2018-12-11 | Wind turbine generator bearing fault diagnosis method based on AVMD and spectrum correlation analysis |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109655266B (en) |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111060301A (en) * | 2019-12-27 | 2020-04-24 | 中国联合网络通信集团有限公司 | Fault diagnosis method and device |
| CN112113766A (en) * | 2020-09-01 | 2020-12-22 | 兰州理工大学 | Characteristic extraction method for early damage state of rolling bearing |
| CN112798280A (en) * | 2021-02-05 | 2021-05-14 | 山东大学 | Rolling bearing fault diagnosis method and system |
| CN112835104A (en) * | 2021-03-26 | 2021-05-25 | 中国石油大学(华东) | Unconventional reservoir natural frequency in-situ measurement system |
| US11539317B2 (en) | 2021-04-05 | 2022-12-27 | General Electric Renovables Espana, S.L. | System and method for detecting degradation in wind turbine generator bearings |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107192554A (en) * | 2017-05-19 | 2017-09-22 | 西安理工大学 | A kind of vibrating failure diagnosis method of Wind turbines rolling bearing |
| CN107832525A (en) * | 2017-11-07 | 2018-03-23 | 昆明理工大学 | A kind of method and its application of information entropy optimization VMD extractions bearing fault characteristics frequency |
| CN108387373A (en) * | 2017-12-06 | 2018-08-10 | 上海电力学院 | The Fault Diagnosis of Roller Bearings of variation mode decomposition is improved based on related coefficient |
-
2018
- 2018-12-11 CN CN201811507395.9A patent/CN109655266B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107192554A (en) * | 2017-05-19 | 2017-09-22 | 西安理工大学 | A kind of vibrating failure diagnosis method of Wind turbines rolling bearing |
| CN107832525A (en) * | 2017-11-07 | 2018-03-23 | 昆明理工大学 | A kind of method and its application of information entropy optimization VMD extractions bearing fault characteristics frequency |
| CN108387373A (en) * | 2017-12-06 | 2018-08-10 | 上海电力学院 | The Fault Diagnosis of Roller Bearings of variation mode decomposition is improved based on related coefficient |
Non-Patent Citations (4)
| Title |
|---|
| 柏林 等: "盲VMD-Cepstral在轴承故障诊断中的应用", 《振动、测试与诊断》 * |
| 柴慧理 等: "基于提高变分模态分解的齿轮箱复合故障特征提取", 《机械传动》 * |
| 郑近德 等: "广义变分模态分解方法及其在变工况齿轮故障诊断中的应用", 《振动工程学报》 * |
| 齐咏生 等: "基于自适应数学形态学的滚动轴承故障诊断方法", 《大连理工大学学报》 * |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111060301A (en) * | 2019-12-27 | 2020-04-24 | 中国联合网络通信集团有限公司 | Fault diagnosis method and device |
| CN112113766A (en) * | 2020-09-01 | 2020-12-22 | 兰州理工大学 | Characteristic extraction method for early damage state of rolling bearing |
| CN112798280A (en) * | 2021-02-05 | 2021-05-14 | 山东大学 | Rolling bearing fault diagnosis method and system |
| CN112798280B (en) * | 2021-02-05 | 2022-01-04 | 山东大学 | Rolling bearing fault diagnosis method and system |
| CN112835104A (en) * | 2021-03-26 | 2021-05-25 | 中国石油大学(华东) | Unconventional reservoir natural frequency in-situ measurement system |
| US11539317B2 (en) | 2021-04-05 | 2022-12-27 | General Electric Renovables Espana, S.L. | System and method for detecting degradation in wind turbine generator bearings |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109655266B (en) | 2021-12-10 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109655266A (en) | A kind of Wind turbines Method for Bearing Fault Diagnosis based on AVMD and spectral coherence analysis | |
| CN105424359B (en) | A kind of gear and bearing mixed fault feature extracting method based on sparse decomposition | |
| CN109297713B (en) | Steam turbine fault diagnosis method based on stable and non-stable vibration signal characteristic selection | |
| CN105760839A (en) | Bearing fault diagnosis method based on multi-feature manifold learning and support vector machine | |
| Liu et al. | Acoustic emission analysis for wind turbine blade bearing fault detection under time-varying low-speed and heavy blade load conditions | |
| Liu et al. | Wind turbine blade bearing fault diagnosis under fluctuating speed operations via Bayesian augmented Lagrangian analysis | |
| CN106778694A (en) | A kind of gear transmission noises analysis method based on set empirical mode decomposition and SVMs | |
| CN105136454A (en) | Wind turbine gear box fault recognition method | |
| CN104655425A (en) | Bearing fault classification diagnosis method based on sparse representation and LDM (large margin distribution machine) | |
| CN109255333B (en) | Fault hybrid diagnosis method for rolling bearing of large wind turbine generator | |
| CN110617960A (en) | Wind turbine generator gearbox fault diagnosis method and system | |
| CN103048133A (en) | Bayesian network-based rolling bearing fault diagnosis method | |
| CN105065212A (en) | Checking method and system of wind generation sets of wind power plant | |
| Sun et al. | A robust approach of multi-sensor fusion for fault diagnosis using convolution neural network | |
| CN113240022A (en) | Wind power gear box fault detection method of multi-scale single-classification convolutional network | |
| CN107036808A (en) | Gearbox of wind turbine combined failure diagnostic method based on SVMs probability Estimation | |
| CN113673442B (en) | Variable working condition fault detection method based on semi-supervised single classification network | |
| CN112160877A (en) | Fan bearing fault diagnosis method based on SFA and CNN | |
| CN113916535B (en) | Bearing diagnosis method, system, equipment and medium based on time frequency and CNN | |
| CN116956215A (en) | Fault diagnosis method and system for transmission system | |
| CN112729825A (en) | Method for constructing bearing fault diagnosis model based on convolution cyclic neural network | |
| CN112986821A (en) | Fault diagnosis method for broken blade of variable pitch motor rotor of offshore wind turbine generator | |
| Cao et al. | Remaining useful life prediction of wind turbine generator bearing based on EMD with an indicator | |
| CN114136604B (en) | Rotary equipment fault diagnosis method and system based on improved sparse dictionary | |
| Zhou et al. | Three-phase asynchronous motor fault diagnosis using attention mechanism and hybrid CNN-MLP by multi-sensor information |
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 |