CN107478418A - A kind of rotating machinery fault characteristic automatic extraction method - Google Patents
A kind of rotating machinery fault characteristic automatic extraction method Download PDFInfo
- Publication number
- CN107478418A CN107478418A CN201710514497.2A CN201710514497A CN107478418A CN 107478418 A CN107478418 A CN 107478418A CN 201710514497 A CN201710514497 A CN 201710514497A CN 107478418 A CN107478418 A CN 107478418A
- Authority
- CN
- China
- Prior art keywords
- mrow
- msub
- msup
- feature
- mfrac
- 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.)
- Pending
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
Landscapes
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses a kind of rotating machinery fault characteristic automatic extraction method, the inventive method are as follows:The first step, sparseness filtering model is trained as training sample using the fault-signal collected, and obtain weight matrix, the higher-dimension fault signature of characteristic of rotating machines vibration signal is extracted by weight matrix;Second step, the visualization that fault signature is realized in dimension-reduction treatment is carried out to the high dimensional feature of acquisition using t SNE algorithms, thus fault diagnosis can be realized to unknown fault-signal according to the feature of extraction.The present invention carries out intelligent diagnostics by unsupervised learning method to fault-signal, more accurate, reasonable than existing methods.
Description
Technical field
The present invention relates to the fault vibration technology of characteristic of rotating machines vibration signal, especially a kind of rotating machinery fault feature is certainly
Dynamic extractive technique.
Background technology
Vibration signal is the carrier of mechanical breakdown feature, and the vibration signal of plant equipment is analyzed, and extraction failure is special
Sign, then judges that the failure of plant equipment is the common method of mechanical fault diagnosis according to fault signature.The usual work of plant equipment
Make in the working environment of more vibration sources, ambient noise is strong, so the mechanical oscillation signal that scene measures is typically that strong background is made an uproar
Multi -components non-stationary signal under sound, in this case, fault signature is extracted from the mechanical oscillation signal of complexity, so as to divide
Just become extremely difficult from mechanical oscillation signal similar in fault mode.Therefore, in order to improve the precision of mechanical fault diagnosis and
Effect, it is necessary to explore application of the new signal processing method in mechanical fault diagnosis.
With the research of machine learning, neutral net starts to attract the concern of more and more scholars.It can be by hiding
Layer is automatically from signal learning to high dimensional feature, but it still needs substantial amounts of label data.It can be substituted as one kind
The selection of manual designs character representation, non-supervisory feature learning are successfully applied in many on the character representation extracted
The work such as image, recording and audio in.But many non-supervisory feature learning algorithms currently are very difficult, because it
Need the regulations of various parameters.The feature learnt if these parameters are not provided be likely to result in one it is very poor
Accuracy rate of diagnosis.These algorithms include sparse Boltzmann machine, sparse autocoder, sparse coding, independent component analysis
Etc..For example sparse Boltzmann machine of the adjustable parameter of these algorithms just has up to that six kinds of parameters need to adjust, independent element point
Analysis only needs to adjust a parameter, but it sets upper scale portrayal very poor for big input or feature.Ngiam etc. is proposed
A kind of non-supervisory feature learning framework is referred to as sparseness filtering, and it is only absorbed in the openness and neglect studies of Optimization Learning feature
The distribution situation of data, while it is very perfect to the scale portrayal for inputting dimension and only a characteristic parameter needs to adjust, because
This sparseness filtering is readily adjusted and is easily achieved by a few row MATLAB codes.Author is carried out using sparseness filtering simultaneously
Image recognition and Classification of Speech, all generate very perfect effect.
Simplicity and its perfect performance due to sparseness filtering, scientific research personnel propose sparseness filtering to solve rotating machinery
Troubleshooting issue.Learn to be characterized in that high dimensional feature does not accomplish Visualization yet with sparseness filtering, therefore have very much
High dimensional feature is mapped to low-dimensional and represented by a kind of suitable dimensionality reduction instrument of necessary choice.
The content of the invention
For the deficiency in above-mentioned technology, carried automatically based on intelligent trouble diagnosis it is an object of the invention to provide one kind
The method for taking vibration signal fault signature, automatically extracting for fault signature is realized by the unsupervised learning mode of sparseness filtering,
And the visualization of high dimensional feature is realized by t-SNE algorithms.
A kind of rotating machinery fault characteristic automatic extraction method proposed by the present invention comprises the following steps:
(1) sample is collected:Using original vibration signal as training sample, adopted using overlapping sampling method from training sample
Collect Z segment signals, these segmentation composition training setsWhereinIt is that j-th of fragmented packets contains NinIndividual data point, Nin
For representing the input dimension of sparseness filtering, NoutRepresent output dimension.
(2) albefaction:Data setWrite matrix formThen whitening pretreatment is carried out.The mesh of albefaction
Be segmentation is directly reduced correlation while accelerate convergence speed.Albefaction uses the Eigenvalues Decomposition of covariance matrix:
Cov (ST)=EUET (1)
Cov (S in formulaT) it is covariance matrix, E is the orthogonal matrix of characteristic vector, and U is the diagonal matrix of characteristic value.
Then albefaction training set SwIt can be calculated as follows:
Sw=EU-1/2ETS (2)
(3) sparseness filtering is trained:Sparseness filtering model passes through by SwTraining, the weight matrix W obtained afterwards are used to calculate
The local feature of training sample.
(4) local feature is calculated:Training sample xiK segmentation is divided into successively, and wherein K is integer and is equal to N/Nin。K
Individual segmentation composition data setAndThen we are calculated each using the sparseness filtering model trained
The local feature of sample
(5) learning characteristic is obtained:Learning characteristic fiObtained by using the method for average to combine local feature:
(6) dimension of learning characteristic is reduced:Due to the feature f of acquisitioniBe still high dimensional data, using t-SNE algorithms come
Its dimension is reduced, the feature then mapped can show on scatter diagram.
The invention has the advantages that realize that the automatic of fault signature carries by the unsupervised learning mode of sparseness filtering
Take, and the visualization of high dimensional feature is realized by t-SNE algorithms.
Compared with conventional method, the present invention proposes the method being combined based on sparseness filtering and t-SNE algorithms can
Effective extraction fault signature simultaneously realizes visualized operation.So this method may apply to the rotating machinery vibrating system failure
In diagnosis, to analyze the type for causing mechanical breakdown, state-detection and fault diagnosis of system etc. are carried out.
Brief description of the drawings
Fig. 1 is a kind of flow chart of rotating machinery fault Automatic signature extraction of the present invention;
Fig. 2 is sparseness filtering system model figure.
Embodiment
The technical scheme of invention is described in detail below in conjunction with the accompanying drawings:
Further explanation is made with reference to implementation of the accompanying drawing to the present invention.Fig. 1 is flow chart of the method for the present invention, such as Fig. 1 institutes
Show, the algorithm includes following three steps.
Step 1:Original vibration signal is pre-processed, detailed process is as follows:
(1) sample is collected:Using original vibration signal as training sample, adopted using overlapping sampling method from training sample
Collect Z segment signals, these segmentation composition training setsWhereinIt is that j-th of fragmented packets contains NinIndividual data point, Nin
For representing the input dimension of sparseness filtering, NoutRepresent output dimension.
(2) albefaction:Data setWrite matrix formThen whitening pretreatment is carried out.The mesh of albefaction
Be segmentation is directly reduced correlation while accelerate convergence speed.Albefaction uses the Eigenvalues Decomposition of covariance matrix:
Cov (ST)=EUET (1)
Cov (S in formulaT) it is covariance matrix, E is the orthogonal matrix of characteristic vector, and U is the diagonal matrix of characteristic value.So
Albefaction training set S afterwardswIt can be calculated as follows:
Sw=EU-1/2ETS (2)
Step 2:To the albefaction training set S described in step 1wSparseness filtering processing is carried out, detailed process is as follows:
The signal for collection is inputted, the feature for exporting to learn.The signal collected is divided into many identical samples
One training set is formed with thisWhereinIt is that a sample M is number of samples.Sample is by using weights
Matrix W ∈ RN×LIt is mapped to characteristic vectorOn, as shown in Figure 2.For sparseness filtering, nonlinear characteristic is to pass through
Obtained using an activation primitive to calculate, in our experiment, activation primitive uses soft-threshold function:
In formulaRepresent j-th of characteristic value of the i-th row, ε=10-8。
Characteristic valueAn eigenmatrix is formed, we, which first normalize, is each characterized as equal activation value.Specific practice
Be by each feature divided by its all samples two norms:
Then each row are normalized by two norms again, and so they will fall on the unit sphere of two norms
.Specific practice is:
At this moment our cans optimize to the feature that these were normalized.We are punished using L1 norms to constrain
It is openness.For a data set for having M sample, the object function of sparseness filtering is expressed as:
Step 3:Sparseness filtering model described in step 2 is passed through by SwTraining, the weight matrix W obtained afterwards is based on
Calculate the local feature of training sample.
(1) local feature is calculated:Training sample xiK segmentation is divided into successively, and wherein K is integer and is equal to N/Nin。K
Individual segmentation composition data setAndThen we are calculated each using the sparseness filtering model trained
The local feature of sample
(2) learning characteristic is obtained:Learning characteristic fiObtained by using the method for average to combine local feature:
Step 4:To the feature f learnt described in step 3iIt is still high dimensional data, it is reduced using t-SNE algorithms
Dimension, the feature then mapped visable representation can come out on scatter diagram.
T-SNE is a Nonlinear Dimension Reduction technology and the high dimensional data that is particularly suitable for use in carries out visualization behaviour in scatter diagram
Make.This method is by defining joint probability distribution pijTo carry out estimating two feature fiAnd fjBetween similitude two-by-two, specifically do
Method is as follows by symmetrical two conditional probabilities:
In superincumbent equation, gaussian kernel function σiIt need to be arranged such:The randomness of conditional probability is equal to predefined mixed
Random degree.Result is gaussian kernel function σiOptimal value become very little and along with a high packing density in data space, because
This σ for each input objectiOptimal value can be determined by using a simple binary search.Low
Dimension space feature yiAnd yj(i.e. fiAnd fjMappings characteristics) between similitude by using a normalized heavy-tailed kernel function
To estimate.Specifically, feature yiAnd yjBetween Joint Distribution pijIt is single free with one by normalizing student t distributions
Spend to calculate:
The heavy-tailed permission dissmilarity feature of normalized student t distributions is modeled into farther distribution in mapping space,
yiPosition by minimizing Joint Distribution qijAnd pijBetween KL divergences determine:
P and Q is p in formulaijAnd qijMatrix form.By minimizing above formula, fiAnd yiBetween become more and more similar, because
This yiF can be representediFeature.
Described above is only the preferred embodiment of the present invention, it should be pointed out that:For the ordinary skill people of the art
For member, under the premise without departing from the principles of the invention, some improvements and modifications can also be made, these improvements and modifications also should
It is considered as protection scope of the present invention.
Claims (4)
1. a kind of rotating machinery fault characteristic automatic extraction method, it is characterized in that, this method is carried out to original vibration signal first
Pretreatment, then, sparseness filtering processing is carried out to treated vibration signal, to sparseness filtering model by being trained by albefaction
The weight matrix obtained afterwards is used for the local feature for calculating training sample;Dimension is finally reduced using t-SNE algorithms, makes to reflect
The feature penetrated visable representation can come out on scatter diagram.
2. a kind of rotating machinery fault characteristic automatic extraction method according to claim 1, it is characterized in that, it is described to original
Beginning vibration signal is pre-processed specially:
Step 1.1 collects sample:Using original vibration signal as training sample, adopted using overlapping sampling method from training sample
Collect Z segment signals, the signal subsection is formed into training setWhereinIt is that j-th of fragmented packets contains NinIndividual data
Point, NinFor representing the input dimension of sparseness filtering, NoutRepresent output dimension;
Step 1.2 albefaction:By data setWrite matrix formThen whitening pretreatment is carried out;It is described white
Change the Eigenvalues Decomposition that processing procedure uses covariance matrix:
cov(ST)=EUET (1)
Cov (S in formulaT) it is covariance matrix, E is the orthogonal matrix of characteristic vector, and U is the diagonal matrix of characteristic value;Then albefaction
Training set SwIt is calculated as follows:
Sw=EU-1/2ETS (2)
3. a kind of rotating machinery fault characteristic automatic extraction method according to claim 2, it is characterized in that, described is sparse
Filter process is specially:
Step 2.1 trains sparseness filtering:Sparseness filtering model passes through by albefaction training set SwTraining, the weight matrix W obtained afterwards
For calculating the local feature of training sample;
Step 2.2 calculates local feature:Training sample xiK segmentation is divided into successively, and wherein K is integer and is equal to N/Nin, K
Individual segmentation composition data setAndThen we are calculated each using the sparseness filtering model trained
The local feature of sample
Step 2.3 obtains learning characteristic:Learning characteristic fiObtained by using the method for average to combine local feature:
<mrow>
<msup>
<mi>f</mi>
<mi>i</mi>
</msup>
<mo>=</mo>
<msup>
<mrow>
<mo>(</mo>
<mfrac>
<mn>1</mn>
<mi>K</mi>
</mfrac>
<munderover>
<mo>&Sigma;</mo>
<mrow>
<mi>k</mi>
<mo>=</mo>
<mn>1</mn>
</mrow>
<mi>K</mi>
</munderover>
<msubsup>
<mi>f</mi>
<mi>k</mi>
<mi>i</mi>
</msubsup>
<mo>)</mo>
</mrow>
<mi>T</mi>
</msup>
<mo>-</mo>
<mo>-</mo>
<mo>-</mo>
<mrow>
<mo>(</mo>
<mn>7</mn>
<mo>)</mo>
</mrow>
</mrow>
4. a kind of rotating machinery fault characteristic automatic extraction method according to claim 1, it is characterized in that, it is described to use t-
SNE algorithms are specially to reduce dimension:
Define joint probability distribution pijTo carry out estimating two feature fiAnd fjBetween similitude two-by-two, symmetrical two conditional probabilities
It is as follows:
<mrow>
<msub>
<mi>p</mi>
<mrow>
<mi>i</mi>
<mo>|</mo>
<mi>j</mi>
</mrow>
</msub>
<mo>=</mo>
<mfrac>
<mrow>
<mi>exp</mi>
<mrow>
<mo>(</mo>
<mo>-</mo>
<mo>|</mo>
<mo>|</mo>
<msub>
<mi>f</mi>
<mi>i</mi>
</msub>
<mo>-</mo>
<msub>
<mi>f</mi>
<mi>j</mi>
</msub>
<mo>|</mo>
<msup>
<mo>|</mo>
<mn>2</mn>
</msup>
<mo>/</mo>
<mn>2</mn>
<msubsup>
<mi>&sigma;</mi>
<mi>i</mi>
<mn>2</mn>
</msubsup>
<mo>)</mo>
</mrow>
</mrow>
<mrow>
<msub>
<mi>&Sigma;</mi>
<mrow>
<mi>k</mi>
<mo>&NotEqual;</mo>
<mi>l</mi>
</mrow>
</msub>
<mi>exp</mi>
<mrow>
<mo>(</mo>
<mo>-</mo>
<mo>|</mo>
<mo>|</mo>
<msub>
<mi>f</mi>
<mi>k</mi>
</msub>
<mo>-</mo>
<msub>
<mi>f</mi>
<mi>l</mi>
</msub>
<mo>|</mo>
<msup>
<mo>|</mo>
<mn>2</mn>
</msup>
<mo>/</mo>
<mn>2</mn>
<msubsup>
<mi>&sigma;</mi>
<mi>i</mi>
<mn>2</mn>
</msubsup>
<mo>)</mo>
</mrow>
</mrow>
</mfrac>
<mo>-</mo>
<mo>-</mo>
<mo>-</mo>
<mrow>
<mo>(</mo>
<mn>8</mn>
<mo>)</mo>
</mrow>
</mrow>
<mrow>
<msub>
<mi>p</mi>
<mrow>
<mi>i</mi>
<mi>j</mi>
</mrow>
</msub>
<mo>=</mo>
<mfrac>
<mrow>
<msub>
<mi>p</mi>
<mrow>
<mi>i</mi>
<mo>|</mo>
<mi>j</mi>
</mrow>
</msub>
<mo>+</mo>
<msub>
<mi>p</mi>
<mrow>
<mi>j</mi>
<mo>|</mo>
<mi>i</mi>
</mrow>
</msub>
</mrow>
<mrow>
<mn>2</mn>
<mi>N</mi>
</mrow>
</mfrac>
<mo>-</mo>
<mo>-</mo>
<mo>-</mo>
<mrow>
<mo>(</mo>
<mn>9</mn>
<mo>)</mo>
</mrow>
</mrow>
Gaussian kernel function σiThe randomness for being arranged to conditional probability is equal to predefined randomness;
In lower dimensional space, feature y is setiAnd yjAs fiAnd fjMappings characteristics, feature yiAnd yjBetween Joint Distribution pijPass through
Normalization student t distributions calculate with a single-degree-of-freedom:
<mrow>
<msub>
<mi>q</mi>
<mrow>
<mi>i</mi>
<mi>j</mi>
</mrow>
</msub>
<mo>=</mo>
<mfrac>
<msup>
<mrow>
<mo>(</mo>
<mn>1</mn>
<mo>+</mo>
<mo>|</mo>
<mo>|</mo>
<msub>
<mi>y</mi>
<mi>i</mi>
</msub>
<mo>-</mo>
<msub>
<mi>y</mi>
<mi>j</mi>
</msub>
<mo>|</mo>
<msup>
<mo>|</mo>
<mn>2</mn>
</msup>
<mo>)</mo>
</mrow>
<mrow>
<mo>-</mo>
<mn>1</mn>
</mrow>
</msup>
<mrow>
<msub>
<mo>&Sigma;</mo>
<mrow>
<mi>k</mi>
<mo>&NotEqual;</mo>
<mi>l</mi>
</mrow>
</msub>
<msup>
<mrow>
<mo>(</mo>
<mn>1</mn>
<mo>+</mo>
<mo>|</mo>
<mo>|</mo>
<msub>
<mi>y</mi>
<mi>k</mi>
</msub>
<mo>-</mo>
<msub>
<mi>y</mi>
<mi>l</mi>
</msub>
<mo>|</mo>
<msup>
<mo>|</mo>
<mn>2</mn>
</msup>
<mo>)</mo>
</mrow>
<mrow>
<mo>-</mo>
<mn>1</mn>
</mrow>
</msup>
</mrow>
</mfrac>
<mo>-</mo>
<mo>-</mo>
<mo>-</mo>
<mrow>
<mo>(</mo>
<mn>10</mn>
<mo>)</mo>
</mrow>
</mrow>
The heavy-tailed permission dissmilarity feature of normalized student t distributions is modeled into farther distribution, y in mapping spaceiPosition
Put by minimizing Joint Distribution qijAnd pijBetween KL divergences determine:
<mrow>
<mi>K</mi>
<mi>L</mi>
<mrow>
<mo>(</mo>
<mi>P</mi>
<mo>|</mo>
<mo>|</mo>
<mi>Q</mi>
<mo>)</mo>
</mrow>
<mo>=</mo>
<munder>
<mo>&Sigma;</mo>
<mi>i</mi>
</munder>
<munder>
<mo>&Sigma;</mo>
<mi>j</mi>
</munder>
<msub>
<mi>p</mi>
<mrow>
<mi>i</mi>
<mi>j</mi>
</mrow>
</msub>
<mi>l</mi>
<mi>o</mi>
<mi>g</mi>
<mrow>
<mo>(</mo>
<mfrac>
<msub>
<mi>p</mi>
<mrow>
<mi>i</mi>
<mi>j</mi>
</mrow>
</msub>
<msub>
<mi>q</mi>
<mrow>
<mi>i</mi>
<mi>j</mi>
</mrow>
</msub>
</mfrac>
<mo>)</mo>
</mrow>
<mo>-</mo>
<mo>-</mo>
<mo>-</mo>
<mrow>
<mo>(</mo>
<mn>11</mn>
<mo>)</mo>
</mrow>
</mrow>
P and Q is p in formulaijAnd qijMatrix form, by minimizing above formula, fiAnd yiBetween become more and more similar, therefore yi
F can be representediFeature.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710514497.2A CN107478418A (en) | 2017-06-29 | 2017-06-29 | A kind of rotating machinery fault characteristic automatic extraction method |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710514497.2A CN107478418A (en) | 2017-06-29 | 2017-06-29 | A kind of rotating machinery fault characteristic automatic extraction method |
Publications (1)
Publication Number | Publication Date |
---|---|
CN107478418A true CN107478418A (en) | 2017-12-15 |
Family
ID=60596101
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201710514497.2A Pending CN107478418A (en) | 2017-06-29 | 2017-06-29 | A kind of rotating machinery fault characteristic automatic extraction method |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN107478418A (en) |
Cited By (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108509701A (en) * | 2018-03-21 | 2018-09-07 | 南京航空航天大学 | A kind of direct intelligent diagnosing method of rotating machinery fault based on vibration signal |
CN108549741A (en) * | 2018-03-07 | 2018-09-18 | 上海交通大学 | Main shaft seals of shield machine performance health evaluating method and system |
CN108827605A (en) * | 2018-03-20 | 2018-11-16 | 南京航空航天大学 | A kind of mechanical breakdown characteristic automatic extraction method based on improvement sparseness filtering |
CN108844733A (en) * | 2018-04-18 | 2018-11-20 | 苏州微著设备诊断技术有限公司 | A kind of gear condition monitoring index extracting method based on KL divergence and root-mean-square value |
CN109784284A (en) * | 2019-01-22 | 2019-05-21 | 上海瀚挚电子科技有限公司 | The self study recognition methods of working condition based on rotating machinery |
CN113505830A (en) * | 2021-07-09 | 2021-10-15 | 西安交通大学 | Rotating machine fault diagnosis method, system, equipment and storage medium |
Citations (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103953490A (en) * | 2014-04-23 | 2014-07-30 | 浙江工业大学 | Implementation method for monitoring status of hydraulic turbine set based on HLSNE |
US9142012B2 (en) * | 2012-05-31 | 2015-09-22 | Apple Inc. | Systems and methods for chroma noise reduction |
CN105809124A (en) * | 2016-03-06 | 2016-07-27 | 北京工业大学 | DWT- and Parametric t-SNE-based characteristic extracting method of motor imagery EEG(Electroencephalogram) signals |
CN106017876A (en) * | 2016-05-11 | 2016-10-12 | 西安交通大学 | Wheel set bearing fault diagnosis method based on equally-weighted local feature sparse filter network |
CN106067042A (en) * | 2016-06-13 | 2016-11-02 | 西安电子科技大学 | Polarization SAR sorting technique based on semi-supervised degree of depth sparseness filtering network |
CN106094786A (en) * | 2016-05-30 | 2016-11-09 | 宁波大学 | Industrial process flexible measurement method based on integrated-type independent entry regression model |
CN106444665A (en) * | 2016-09-22 | 2017-02-22 | 宁波大学 | Fault classification diagnosis method based on non-Gaussian similarity matching |
-
2017
- 2017-06-29 CN CN201710514497.2A patent/CN107478418A/en active Pending
Patent Citations (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US9142012B2 (en) * | 2012-05-31 | 2015-09-22 | Apple Inc. | Systems and methods for chroma noise reduction |
CN103953490A (en) * | 2014-04-23 | 2014-07-30 | 浙江工业大学 | Implementation method for monitoring status of hydraulic turbine set based on HLSNE |
CN105809124A (en) * | 2016-03-06 | 2016-07-27 | 北京工业大学 | DWT- and Parametric t-SNE-based characteristic extracting method of motor imagery EEG(Electroencephalogram) signals |
CN106017876A (en) * | 2016-05-11 | 2016-10-12 | 西安交通大学 | Wheel set bearing fault diagnosis method based on equally-weighted local feature sparse filter network |
CN106094786A (en) * | 2016-05-30 | 2016-11-09 | 宁波大学 | Industrial process flexible measurement method based on integrated-type independent entry regression model |
CN106067042A (en) * | 2016-06-13 | 2016-11-02 | 西安电子科技大学 | Polarization SAR sorting technique based on semi-supervised degree of depth sparseness filtering network |
CN106444665A (en) * | 2016-09-22 | 2017-02-22 | 宁波大学 | Fault classification diagnosis method based on non-Gaussian similarity matching |
Non-Patent Citations (1)
Title |
---|
谷玉海: "大型风电机组齿轮箱早期故障诊断技术与系统研究", 《中国博士学位论文全文数据库工程科技Ⅱ辑》 * |
Cited By (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108549741A (en) * | 2018-03-07 | 2018-09-18 | 上海交通大学 | Main shaft seals of shield machine performance health evaluating method and system |
CN108827605A (en) * | 2018-03-20 | 2018-11-16 | 南京航空航天大学 | A kind of mechanical breakdown characteristic automatic extraction method based on improvement sparseness filtering |
CN108509701A (en) * | 2018-03-21 | 2018-09-07 | 南京航空航天大学 | A kind of direct intelligent diagnosing method of rotating machinery fault based on vibration signal |
CN108509701B (en) * | 2018-03-21 | 2021-08-06 | 南京航空航天大学 | Rotary machine fault direct intelligent diagnosis method based on vibration signal |
CN108844733A (en) * | 2018-04-18 | 2018-11-20 | 苏州微著设备诊断技术有限公司 | A kind of gear condition monitoring index extracting method based on KL divergence and root-mean-square value |
CN109784284A (en) * | 2019-01-22 | 2019-05-21 | 上海瀚挚电子科技有限公司 | The self study recognition methods of working condition based on rotating machinery |
CN113505830A (en) * | 2021-07-09 | 2021-10-15 | 西安交通大学 | Rotating machine fault diagnosis method, system, equipment and storage medium |
CN113505830B (en) * | 2021-07-09 | 2024-04-02 | 西安交通大学 | Rotary machine fault diagnosis method, system, equipment and storage medium |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN107478418A (en) | A kind of rotating machinery fault characteristic automatic extraction method | |
Chen et al. | Integrating the data augmentation scheme with various classifiers for acoustic scene modeling | |
Li et al. | Mechanical fault diagnosis based on redundant second generation wavelet packet transform, neighborhood rough set and support vector machine | |
WO2019090879A1 (en) | Analog circuit fault diagnosis method based on cross wavelet features | |
CN103728551B (en) | A kind of analog-circuit fault diagnosis method based on cascade integrated classifier | |
CN105841961A (en) | Bearing fault diagnosis method based on Morlet wavelet transformation and convolutional neural network | |
CN105760839A (en) | Bearing fault diagnosis method based on multi-feature manifold learning and support vector machine | |
CN108827605A (en) | A kind of mechanical breakdown characteristic automatic extraction method based on improvement sparseness filtering | |
CN110017991A (en) | Rolling bearing fault classification method and system based on spectrum kurtosis and neural network | |
CN110046562A (en) | A kind of wind power system health monitor method and device | |
CN105738109A (en) | Bearing fault classification diagnosis method based on sparse representation and ensemble learning | |
CN106653032A (en) | Animal sound detecting method based on multiband energy distribution in low signal-to-noise-ratio environment | |
CN101750210A (en) | Fault diagnosis method based on OLPP feature reduction | |
CN109323754A (en) | A kind of train wheel polygon fault diagnosis detection method | |
CN111341294B (en) | Method for converting text into voice with specified style | |
CN105678343A (en) | Adaptive-weighted-group-sparse-representation-based diagnosis method for noise abnormity of hydroelectric generating set | |
CN102539159A (en) | Fault diagnosis method for valve mechanism of diesel engine | |
CN105609116A (en) | Speech emotional dimensions region automatic recognition method | |
CN101650944A (en) | Method for distinguishing speakers based on protective kernel Fisher distinguishing method | |
CN107133640A (en) | Image classification method based on topography's block description and Fei Sheer vectors | |
CN108509701A (en) | A kind of direct intelligent diagnosing method of rotating machinery fault based on vibration signal | |
CN110020637A (en) | A kind of analog circuit intermittent fault diagnostic method based on more granularities cascade forest | |
CN106295708A (en) | A kind of continuous data preprocess method based on Fisher classifier group | |
CN113327632A (en) | Unsupervised abnormal sound detection method and unsupervised abnormal sound detection device based on dictionary learning | |
Zhang et al. | Temporal Transformer Networks for Acoustic Scene Classification. |
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 | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20171215 |
|
RJ01 | Rejection of invention patent application after publication |