CN106443447A - An aero-generator fault feature extraction method based on iSDAE - Google Patents
An aero-generator fault feature extraction method based on iSDAE Download PDFInfo
- Publication number
- CN106443447A CN106443447A CN201610871333.0A CN201610871333A CN106443447A CN 106443447 A CN106443447 A CN 106443447A CN 201610871333 A CN201610871333 A CN 201610871333A CN 106443447 A CN106443447 A CN 106443447A
- Authority
- CN
- China
- Prior art keywords
- noise reduction
- fault
- autocoder
- training
- aerogenerator
- 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
- G01R—MEASURING ELECTRIC VARIABLES; MEASURING MAGNETIC VARIABLES
- G01R31/00—Arrangements for testing electric properties; Arrangements for locating electric faults; Arrangements for electrical testing characterised by what is being tested not provided for elsewhere
- G01R31/34—Testing dynamo-electric machines
Abstract
The invention discloses an aero-generator fault feature extraction method based on iSDAE (improved Stacked Denoising Auto-Encoderes), and is mainly to solve the problem of not high diagnosis accuracy since the existing fault diagnosis technology is limited by artificial feature extraction. The method has the following specific steps: 1) fault analysis; 2) data acquisition; 3) data pre-processing; 4) training of the iSDAE (improved Stacked Denoising Auto-Encoderes); and 5) feature output. The method can learn data features automatically and obtains distributed feature representations of original data, has a certain noise immunity and good robustness and effectively improves aero-generator fault diagnosis correct rate.
Description
Technical field
The present invention relates to a kind of, based on iSDAE, (improved Stacked Denoising Autoencoders improves
Storehouse noise reduction autocoder) aerogenerator fault signature extracting method, belong to generator state monitoring and fault diagnosis
Technical field.
Background technology
Aerogenerator is the important component part of main aircraft power source, it be responsible for instrument on aircraft, instrument, radar,
On illumination, radio communication and machine, various control systems etc. provide power supply.Any one link of aerogenerator breaks down, no
Only can be affected it normally run, be possible to lead to aircraft to be unable to normal flight simultaneously, when serious, even can cause great aviation
Accident.Thus carry out the research of aerogenerator fault diagnosis technology in a deep going way, the potential faults that aerogenerator is likely to occur
Make and in time, accurately and rapidly judging it is ensured that the safe operation of aircraft, there is extremely important realistic meaning and huge
Economic benefit.
Now widely used aboard is rotating-rectifier type three-stage brushless synchronous generator, and it is mainly by secondary excitation
Machine, AC exciter and three parts of main generator form, and wherein, pilot exciter is rotary magnetic pole type permanent magnet generator, exchange
Exciter is revolving-armature type synchronous generator, and main generator is rotary pole formula synchronous generator.The structure of aerogenerator
Sufficiently complex, failure mode is various, by carrying out fault mode, impact and HAZAN to aerogenerator, determines aviation
The main fault mode of electromotor has faults in rotating rectifiers, machines under rotor winding faults, stator winding faults, rotating shaft fault and axle
Hold fault etc., wherein, each fault mode can be divided into different fault types again, and such as faults in rotating rectifiers can be divided into again
Single tube fault, two-tube fault etc., bearing fault can be divided into spot corrosion, crackle etc. again.
In fault diagnosis field, have method based on model for the main diagnostic method of these faults, based on signal at
The method of reason and the method based on artificial intelligence, the common practice adopting at present is that signal processing is combined with artificial intelligence,
Particularly, typically all first gather fault-signal, then signal processing is carried out to the fault-signal collecting, artificial extraction event
The fault signature of extraction finally is used for training grader, thus carrying out failure modes by barrier feature.But current feature extraction
Method generally depends on and manually extracts, time and effort consuming, is affected larger by noise jamming, and does not have universality, for
Limited by artificial extraction feature in existing fault diagnosis technology and led to the not high problem of rate of correct diagnosis, the present invention is proposed
A kind of aerogenerator fault signature extracting method based on iSDAE, the method automatically can carry out data characteristicses study, obtain
Distributed nature to initial data represents, and has certain noise resisting ability, has good robustness, effectively improves
Aerogenerator fault diagnosis accuracy.
Content of the invention
The present invention proposes a kind of aerogenerator fault signature extracting method based on iSDAE, and the method is applied to
Generator state monitoring and fault diagnosis field, can automatically carry out data characteristicses study, obtain the distributed of initial data
Character representation, and have certain noise resisting ability, there is good robustness, effectively improve aerogenerator fault diagnosis
Accuracy.
The present invention for achieving the above object, adopts the following technical scheme that：
A kind of aviation alternator rotating rectifier on-line fault diagnosis method based on iSDAE, comprises the steps：
(1) accident analysis.Fault mode, impact and HAZAN are carried out on aerogenerator, determines aerogenerator
Chife failure models and required collection diagnostic signal.Through analysis, aerogenerator mainly has faults in rotating rectifiers, turns
The fault modes such as sub- winding failure, stator winding faults, rotating shaft and bearing fault, the diagnostic signal that need to gather is that main generator is defeated
Go out voltage signal, AC exciter exciting current signal, fuselage shaking signal and rotating shaft torsion signal.
(2) data acquisition.Fault simulation experiment is carried out on generator failure simulation experiment platform, by institute in step (1)
State the diagnostic signal of four kinds of need collections, respectively through voltage sensor, current sensor, vibrating sensor and torque sensor,
Connected with data collecting card again and carry out data acquisition to computer.
(3) data prediction.The dimension of four kinds of diagnostic signals due to collecting in step (2) is different, in order that signal
There is unified statistical distribution, four kinds of signals are normalized, then by four in the case of each fault type
Plant signal group and become column vector, generate sample.
(4) training improves storehouse noise reduction autocoder.The sample obtaining in above-mentioned steps (3) is passed through unsupervised side
Formula trains storehouse noise reduction autocoder, and the distributed nature of study initial data represents.
(4.1) series of noise reduction autocoder needed for setting simultaneously carries out noise reduction autocoder weights and biases initial
Change.Noise reduction autocoder is a kind of neural networks with single hidden layer, because traditional noise reduction autocoder adopts random initializtion
Network weight and the method for biasing, impact to whole autocoder performance, and the present invention adopts fruit bat optimized algorithm
(Fruit Fly Optimization Algorithm, FFOA) first encodes to network weight and biasing, and search obtains one
Then this network parameter is trained by individual more excellent solution as the initial parameter of autocoder again, finally trains optimum
Network parameter.
(4.1-a) initialize.Noise reduction autocoder weights and biasing are encoded, determines the rule of initial fruit bat population
Mould, maximum iteration time, and the initial position of fruit bat population is initialized.
(4.1-b) olfactory sensation random search.Make the primary iteration number of times g=0 of fruit bat algorithm, set fruit bat in iterative process
Body is looked for food random flight direction rand () in stage and arbitrary width in olfactory sensation.
(4.1-c) determine flavor concentration decision content, and calculate the individual odorousness value of fruit bat, now, by reality output
Error E between value and exact value is as taste decision function.
(4.1-d) vision localization.Seek the minimum individuality of odorousness (i.e. error E) as optimum individual, and record this
When individual position and flavor concentration, meanwhile, whole fruit bat colony is flown to optimal location using sharp vision.
(4.1-e) iteration optimizing.Judge whether to reach end condition, that is, whether iterationses reach maximum iteration time.
If meeting, terminating algorithm, if being unsatisfactory for, continuing repeat step (4.1-b) to step (4.1-e), circulating this process.Until repeatedly
When generation number reaches maximum iteration time, terminate algorithm.
(4.2) noise reduction autocoder training.Input training sample, and manually add additive Gaussian noise in the sample,
Make autocoder have certain noise resisting ability, calculate autocoder output.Because autocoder is using unsupervised
Training method it is desirable to obtain and input identical output, reconstructed error is obtained according to input and output, constantly adjustment weights and
Biasing is so that reconstructed error is minimum.
(4.3), after completing the training of one-level noise reduction autocoder, preserve the weights of coded portion and biasing, now noise reduction from
The hidden layer output of dynamic encoder is the level one data feature being learnt, and using this data characteristics as next stage noise reduction certainly
The training sample of dynamic encoder, repeat step (4.1) to step (4.3), until complete to set the noise reduction autocoder of number
Training, as improves storehouse noise reduction autocoder.
(5) feature output.Above-mentioned steps complete the distributed nature study of initial data, remain the spy of initial data
Reference ceases, can by study to feature input and carry out failure modes to grader.
The present invention has the beneficial effect that：.
The present invention proposes a kind of aerogenerator fault signature extracting method based on iSDAE, and the method is applied to
Generator state monitoring and fault diagnosis field, can automatically carry out data characteristicses study, obtain the distributed nature of data
Represent, and have certain noise resisting ability, there is good robustness, effectively improve aerogenerator fault diagnosis correct
Rate.
Brief description
Fig. 1 feature extraction flow chart
Fig. 2 iSDAE network structure
Specific embodiment
Below in conjunction with the accompanying drawings technical scheme is described in detail.As shown in figure 1, a kind of based on iSDAE's
Aerogenerator fault signature extracting method specific embodiment is as follows：
(1) accident analysis.Fault mode, impact and HAZAN are carried out on aerogenerator, determines aerogenerator
Fault mode and required collection diagnostic signal.Through analysis, aerogenerator mainly have faults in rotating rectifiers, rotor around
The fault modes such as group fault, stator winding faults, rotating shaft and bearing fault, the diagnostic signal that need to gather exports electricity for main generator
Pressure signal, AC exciter exciting current signal, fuselage shaking signal and rotating shaft torsion signal.
(2) data acquisition.Fault simulation experiment is carried out on generator failure simulation experiment platform, by institute in step (1)
State the diagnostic signal of four kinds of need collections, respectively through voltage sensor, current sensor, vibrating sensor and torque sensor,
Connected with data collecting card again and carry out data acquisition to computer.
(3) data prediction.The dimension of four kinds of diagnostic signals due to collecting in step (2) is different, in order that signal
There is unified statistical distribution, four kinds of signals are normalized, specific normalization formula is as follows：
Wherein：XnewFor the fault-signal after normalization, X is the fault-signal before normalization, XmeanFor sample average, Xstd
Standard deviation for sample.
Then continuous for four kinds of signals in the case of each fault type 200 points are formed 1 column vector yn, each
Column vector, as a sample, all samples is formed a sample set Y={ y1, y2, y3...yn...yN, N number of sample altogether.
(4) training improves storehouse noise reduction autocoder.The sample obtaining in above-mentioned steps (3) is passed through unsupervised side
Formula trains storehouse noise reduction autocoder, and the distributed nature of study initial data represents.As shown in Fig. 2 concrete training step
As follows：
(4.1) series of noise reduction autocoder needed for setting simultaneously carries out noise reduction autocoder weights and biases initial
Change.The present invention sets the series of noise reduction autocoder as 4 grades, and the neuron number of first order noise reduction autocoder is 800-
600-800, the neuron number of the second level is 600-400-600, and the neuron number of the third level is 400-200-400, the 4th
The neuron number of level is 200-100-200, and the feature finally learning is the hidden layer of afterbody noise reduction autocoder
Output vector, the neuron activation functions being adopted are sigmoid function, and formula is as follows：
Wherein, e is natural constant.
Noise reduction autocoder is a kind of neural networks with single hidden layer, because traditional noise reduction autocoder is using at random just
Beginningization network weight and the method for biasing, impact to whole autocoder performance, and the present invention adopts fruit bat optimized algorithm
First network weight and biasing are encoded, search obtains a more excellent solution, then using this network parameter as autocoder
Initial parameter be trained again, finally train optimum network parameter.Specific fruit bat optimized algorithm optimizes weights and threshold
Value step is as follows：
(4.1-a) initialize.Noise reduction autocoder weights and biasing are encoded, determines the rule of initial fruit bat population
Mould, maximum iteration time, and the initial position of fruit bat population is initialized.
(4.1-b) olfactory sensation random search.Make the primary iteration number of times g=0 of fruit bat algorithm, set fruit bat in iterative process
Body is looked for food random flight direction rand () in stage and arbitrary width in olfactory sensation.
(4.1-c) determine flavor concentration decision content, and calculate the individual odorousness value of fruit bat, now, by reality output
Error E between value and exact value is as taste decision function.
(4.1-d) vision localization.Seek the minimum individuality of odorousness (i.e. error E) as optimum individual, and record this
When individual position and flavor concentration, meanwhile, whole fruit bat colony is flown to optimal location using sharp vision.
(4.1-e) iteration optimizing.Judge whether to reach end condition, that is, whether iterationses reach maximum iteration time.
If meeting, terminating algorithm, if being unsatisfactory for, continuing repeat step (4.1-b) to step (4.1-e), circulating this process.Until repeatedly
When generation number reaches maximum iteration time, terminate algorithm.
(4.2) noise reduction autocoder training.Input training sample, and manually add Gaussian noise in the sample, make certainly
Dynamic encoder has certain noise resisting ability, calculates autocoder output.Because autocoder adopts unsupervised instruction
The mode of white silk, it is desirable to obtain and input identical output, obtains reconstructed error according to input and output.Constantly adjustment weights and partially
Put so that reconstructed error is minimum.Comprise the following steps that：
(4.2-a) and set allowable error ε and learning rate α, then carry out DAE network training.Input N number of training sample
This, calculate the output of noise reduction autocoder.
(4.2-b) due to noise reduction autocoder using unsupervised training method it is desirable to obtain with input identical defeated
Go out, reconstructed error is obtained according to input and output, reconstructed error formula is
Wherein, Y represents training sample, hW, b(Y) represent the output valve through network calculations for the training sample.
(4.2-c) according to reconstructed error, weight and biasing are adjusted, concrete formula is as follows：
Wherein, WijRepresent the network weight of i-th j-th neuron of layer network, biRepresent i-th layer of biasing,Represent that (W, b) to W for JijSeek local derviation,Represent that (W, b) to b for JiSeek local derviation, l represents iteration time
Number.
(4.2-d) whether decision errors meet allowable error ε and require or whether reach iterationses, failing to meet will
Then repeat step (4.2-b) is asked to step (4.2-d), to require or reach iterationses to tie until whole network output meets expectation
Shu Xunhuan.
(4.3), after completing the training of one-level noise reduction autocoder, preserve the weights of coded portion and biasing, now noise reduction from
The hidden layer output of dynamic encoder is the level one data feature being learnt, and using this data characteristics as next stage noise reduction certainly
The training sample of dynamic encoder, repeat step (4.1) to step (4.3), until complete to set the noise reduction autocoder of number
Training, as improves storehouse noise reduction autocoder.
(5) feature output.Above-mentioned steps complete the distributed nature study of initial data, remain the spy of initial data
Reference ceases, can by study to feature input and carry out failure modes to grader.
Claims (3)
1. a kind of aerogenerator fault signature extracting method based on iSDAE is it is characterised in that comprise the steps of：
Step one：Accident analysis.Fault mode, impact and HAZAN are carried out on aerogenerator, determines aerogenerator
Chife failure models and required collection diagnostic signal.Through analysis, aerogenerator mainly has faults in rotating rectifiers, turns
The fault modes such as sub- winding failure, stator winding faults, rotating shaft and bearing fault, the diagnostic signal that need to gather is that main generator is defeated
Go out voltage signal, AC exciter exciting current signal, fuselage shaking signal and rotating shaft torsion signal.
Step 2：Data acquisition.Generator failure simulation experiment platform carries out fault simulation experiment, described in step one
The diagnostic signal of four kinds of need collections, respectively through voltage sensor, current sensor, vibrating sensor and torque sensor, then
Connected with data collecting card and carry out data acquisition to computer.
Step 3：Data prediction.The dimension of four kinds of diagnostic signals due to collecting in step 2 is different, in order that signal tool
There is unified statistical distribution, four kinds of signals are normalized, then by four kinds in the case of each fault type
Signal group becomes column vector, generates sample.
Step 4：Training improves storehouse noise reduction autocoder.The sample obtaining in above-mentioned steps four is passed through unsupervised side
Formula trains storehouse noise reduction autocoder, and the distributed nature of study initial data represents.
Step 5：Feature exports.Above-mentioned steps complete the distributed nature study of initial data, remain the spy of initial data
Reference ceases, can by study to feature input and carry out failure modes to grader.
2. a kind of aerogenerator fault signature extracting method based on iSDAE according to claim 1, its feature exists
In comprising the following steps that of, the pre-training noise reduction autocoder described in step 4：
Step one：The series of noise reduction autocoder needed for setting simultaneously carries out noise reduction autocoder weights and biasing initialization.
Step 2：Noise reduction autocoder is trained.Input sample simultaneously manually adds additive Gaussian noise in the sample, makes automatically to compile
Code utensil has certain noise resisting ability, calculates autocoder output.Because autocoder adopts unsupervised training side
Formula, it is desirable to obtain and input identical output, obtains reconstructed error according to input and output, constantly adjustment weights and biasing, make
Obtain reconstructed error minimum.
Step 3：After completing the training of one-level noise reduction autocoder, preserve weights and the biasing of coded portion, now noise reduction is automatic
The hidden layer output of encoder is the level one data feature being learnt, and this data characteristics are automatic as next stage noise reduction
The training sample of encoder, repeat step one to step 3, until completing to set the noise reduction autocoder training of series, that is, complete
Storehouse noise reduction autocoder has been become to train.
3. a kind of aerogenerator Fault Diagnosis of Rotating Rectifier method based on iSDAE according to claim 1, it is special
Levy and be, the weights of pre-training noise reduction autocoder described in step 4 and the initialized step of biasing are as follows：
Step one：Initialization.The weights and biasing of noise reduction autocoder are encoded, determines the rule of initial fruit bat population
Mould, maximum iteration time, and the initial position of fruit bat population is initialized.
Step 2：Olfactory sensation random search.The primary iteration number of times making fruit bat algorithm is zero, sets fruit bat individuality in iterative process and exists
Olfactory sensation is looked for food the random flight direction in stage and arbitrary width.
Step 3：Determine flavor concentration decision content, and calculate the individual odorousness value of fruit bat, now, by real output value with
Error between exact value is as taste decision function.
Step 4：Vision localization.Seek the minimum individuality of odorousness as optimum individual, and record now individual position with
And flavor concentration, meanwhile, whole fruit bat colony is flown to optimal location using sharp vision.
Step 5：Iteration optimizing.Judge whether to reach end condition, that is, whether iterationses reach maximum iteration time.If full
Sufficient then terminate algorithm, if being unsatisfactory for, continuing repeat step two to step 5, circulating this process.Until iterationses reach maximum
During iterationses, terminate algorithm.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610871333.0A CN106443447B (en) | 2016-09-26 | 2016-09-26 | A kind of aerogenerator fault signature extracting method based on iSDAE |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610871333.0A CN106443447B (en) | 2016-09-26 | 2016-09-26 | A kind of aerogenerator fault signature extracting method based on iSDAE |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN106443447A true CN106443447A (en) | 2017-02-22 |
| CN106443447B CN106443447B (en) | 2019-05-21 |
Family
ID=58171685
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610871333.0A Active CN106443447B (en) | 2016-09-26 | 2016-09-26 | A kind of aerogenerator fault signature extracting method based on iSDAE |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106443447B (en) |
Cited By (14)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107247231A (en) * | 2017-07-28 | 2017-10-13 | 南京航空航天大学 | A kind of aerogenerator fault signature extracting method based on OBLGWO DBN models |
| CN107316046A (en) * | 2017-03-09 | 2017-11-03 | 河北工业大学 | A kind of method for diagnosing faults that Dynamic adaptiveenhancement is compensated based on increment |
| CN107316087A (en) * | 2017-07-03 | 2017-11-03 | 中国航空工业集团公司西安飞机设计研究所 | It is a kind of to judge the method that aeronautical product tape jam is used |
| CN108241298A (en) * | 2018-01-09 | 2018-07-03 | 南京航空航天大学 | A kind of aerogenerator method for diagnosing faults based on FWA-RNN models |
| CN109190304A (en) * | 2018-10-16 | 2019-01-11 | 南京航空航天大学 | Gas path component fault signature extracts and fault recognition method in a kind of aero-engine whole envelope |
| CN109598336A (en) * | 2018-12-05 | 2019-04-09 | 国网江西省电力有限公司信息通信分公司 | A kind of Data Reduction method encoding neural network certainly based on stack noise reduction |
| CN109599872A (en) * | 2018-12-29 | 2019-04-09 | 重庆大学 | Probabilistic optimal load flow calculation method based on storehouse noise reduction autocoder |
| CN109613428A (en) * | 2018-12-12 | 2019-04-12 | 广州汇数信息科技有限公司 | It is a kind of can be as system and its application in motor device fault detection method |
| CN109858345A (en) * | 2018-12-25 | 2019-06-07 | 华中科技大学 | A kind of intelligent failure diagnosis method suitable for pipe expanding equipment |
| CN110018417A (en) * | 2019-05-24 | 2019-07-16 | 湖南大学 | Method of Motor Fault Diagnosis, system and medium based on the detection of radial stray flux |
| CN110261773A (en) * | 2019-07-01 | 2019-09-20 | 南京航空航天大学 | A kind of aerogenerator failure symptom extracting method and system |
| CN112434298A (en) * | 2021-01-26 | 2021-03-02 | 浙江大学 | Network threat detection system based on self-encoder integration |
| CN113049035A (en) * | 2021-03-12 | 2021-06-29 | 辽宁工程技术大学 | Transformer state monitoring system based on Internet of things |
| CN113625164A (en) * | 2021-08-02 | 2021-11-09 | 南京航空航天大学 | Aviation generator fault feature extraction method, system, medium and computing device |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20150106794A1 (en) * | 2013-10-14 | 2015-04-16 | Nec Laboratories America, Inc. | Transparent performance inference of whole software layers and context-sensitive performance debugging |
| CN104748962A (en) * | 2015-04-03 | 2015-07-01 | 西安交通大学 | Planetary gear box intelligent diagnosis method based on stacking automatic encoding machine |
| CN105184368A (en) * | 2015-09-07 | 2015-12-23 | 中国科学院深圳先进技术研究院 | Distributed extreme learning machine optimization integrated framework system and method |
| CN105957092A (en) * | 2016-05-31 | 2016-09-21 | 福州大学 | Mammary gland molybdenum target image feature self-learning extraction method for computer-aided diagnosis |
-
2016
- 2016-09-26 CN CN201610871333.0A patent/CN106443447B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20150106794A1 (en) * | 2013-10-14 | 2015-04-16 | Nec Laboratories America, Inc. | Transparent performance inference of whole software layers and context-sensitive performance debugging |
| CN104748962A (en) * | 2015-04-03 | 2015-07-01 | 西安交通大学 | Planetary gear box intelligent diagnosis method based on stacking automatic encoding machine |
| CN105184368A (en) * | 2015-09-07 | 2015-12-23 | 中国科学院深圳先进技术研究院 | Distributed extreme learning machine optimization integrated framework system and method |
| CN105957092A (en) * | 2016-05-31 | 2016-09-21 | 福州大学 | Mammary gland molybdenum target image feature self-learning extraction method for computer-aided diagnosis |
Non-Patent Citations (1)
| Title |
|---|
| 汪彩霞等: "基于堆栈降噪自动编码模型的动态纹理分类方法", 《现代电子技术》 * |
Cited By (21)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107316046B (en) * | 2017-03-09 | 2020-08-25 | 河北工业大学 | Fault diagnosis method based on incremental compensation dynamic self-adaptive enhancement |
| CN107316046A (en) * | 2017-03-09 | 2017-11-03 | 河北工业大学 | A kind of method for diagnosing faults that Dynamic adaptiveenhancement is compensated based on increment |
| CN107316087A (en) * | 2017-07-03 | 2017-11-03 | 中国航空工业集团公司西安飞机设计研究所 | It is a kind of to judge the method that aeronautical product tape jam is used |
| CN107316087B (en) * | 2017-07-03 | 2021-08-17 | 中国航空工业集团公司西安飞机设计研究所 | Method for judging fault use of aviation product |
| CN107247231A (en) * | 2017-07-28 | 2017-10-13 | 南京航空航天大学 | A kind of aerogenerator fault signature extracting method based on OBLGWO DBN models |
| CN108241298A (en) * | 2018-01-09 | 2018-07-03 | 南京航空航天大学 | A kind of aerogenerator method for diagnosing faults based on FWA-RNN models |
| CN109190304A (en) * | 2018-10-16 | 2019-01-11 | 南京航空航天大学 | Gas path component fault signature extracts and fault recognition method in a kind of aero-engine whole envelope |
| CN109598336A (en) * | 2018-12-05 | 2019-04-09 | 国网江西省电力有限公司信息通信分公司 | A kind of Data Reduction method encoding neural network certainly based on stack noise reduction |
| CN109613428A (en) * | 2018-12-12 | 2019-04-12 | 广州汇数信息科技有限公司 | It is a kind of can be as system and its application in motor device fault detection method |
| CN109858345B (en) * | 2018-12-25 | 2021-06-11 | 华中科技大学 | Intelligent fault diagnosis method suitable for pipe expansion equipment |
| CN109858345A (en) * | 2018-12-25 | 2019-06-07 | 华中科技大学 | A kind of intelligent failure diagnosis method suitable for pipe expanding equipment |
| CN109599872A (en) * | 2018-12-29 | 2019-04-09 | 重庆大学 | Probabilistic optimal load flow calculation method based on storehouse noise reduction autocoder |
| CN109599872B (en) * | 2018-12-29 | 2022-11-08 | 重庆大学 | Power system probability load flow calculation method based on stack noise reduction automatic encoder |
| CN110018417B (en) * | 2019-05-24 | 2020-05-15 | 湖南大学 | Motor fault diagnosis method, system and medium based on radial stray magnetic flux detection |
| CN110018417A (en) * | 2019-05-24 | 2019-07-16 | 湖南大学 | Method of Motor Fault Diagnosis, system and medium based on the detection of radial stray flux |
| CN110261773A (en) * | 2019-07-01 | 2019-09-20 | 南京航空航天大学 | A kind of aerogenerator failure symptom extracting method and system |
| CN110261773B (en) * | 2019-07-01 | 2021-04-13 | 南京航空航天大学 | Aviation generator fault symptom extraction method and system |
| CN112434298A (en) * | 2021-01-26 | 2021-03-02 | 浙江大学 | Network threat detection system based on self-encoder integration |
| CN113049035A (en) * | 2021-03-12 | 2021-06-29 | 辽宁工程技术大学 | Transformer state monitoring system based on Internet of things |
| CN113049035B (en) * | 2021-03-12 | 2022-05-27 | 辽宁工程技术大学 | Transformer state monitoring system based on Internet of things |
| CN113625164A (en) * | 2021-08-02 | 2021-11-09 | 南京航空航天大学 | Aviation generator fault feature extraction method, system, medium and computing device |
Also Published As
| Publication number | Publication date |
|---|---|
| CN106443447B (en) | 2019-05-21 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106443447A (en) | An aero-generator fault feature extraction method based on iSDAE | |
| WO2023044979A1 (en) | Mechanical fault intelligent diagnosis method under class unbalanced dataset | |
| Zhang et al. | The multiclass fault diagnosis of wind turbine bearing based on multisource signal fusion and deep learning generative model | |
| Rahimilarki et al. | Convolutional neural network fault classification based on time-series analysis for benchmark wind turbine machine | |
| CN109102005B (en) | Small sample deep learning method based on shallow model knowledge migration | |
| CN105354587B (en) | A kind of method for diagnosing faults of wind-driven generator group wheel box | |
| CN108680358A (en) | A kind of Wind turbines failure prediction method based on bearing temperature model | |
| CN106124212A (en) | Based on sparse coding device and the Fault Diagnosis of Roller Bearings of support vector machine | |
| Han et al. | Fault diagnosis system of induction motors based on neural network and genetic algorithm using stator current signals | |
| CN107909118A (en) | A kind of power distribution network operating mode recording sorting technique based on deep neural network | |
| CN104121949A (en) | Condition monitoring method of ship electric propulsion system | |
| CN109597401A (en) | A kind of equipment fault diagnosis method based on data-driven | |
| CN107392304A (en) | A kind of Wind turbines disorder data recognition method and device | |
| CN107247231A (en) | A kind of aerogenerator fault signature extracting method based on OBLGWO DBN models | |
| CN109444740A (en) | A kind of the malfunction intellectual monitoring and diagnostic method of Wind turbines | |
| CN110969194B (en) | Cable early fault positioning method based on improved convolutional neural network | |
| CN105242205A (en) | Aviation three-level AC power generator rotary rectifier online fault diagnosis method | |
| CN107944648A (en) | A kind of accurate Forecasting Methodology of large ship speed of a ship or plane rate of fuel consumption | |
| CN108647786A (en) | The rotating machinery on-line fault monitoring method of neural network is fought based on depth convolution | |
| CN102609764A (en) | CPN neural network-based fault diagnosis method for stream-turbine generator set | |
| CN110417005B (en) | Transient stability serious fault screening method combining deep learning and simulation calculation | |
| CN108154223A (en) | Power distribution network operating mode recording sorting technique based on network topology and long timing information | |
| CN112926728B (en) | Small sample turn-to-turn short circuit fault diagnosis method for permanent magnet synchronous motor | |
| CN110132627A (en) | A kind of method for diagnosing faults of propeller | |
| CN111597996A (en) | Method for constructing wind turbine generator bearing fault identification model based on deep learning |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |