CN112861275A - Rotary machine fault diagnosis method based on minimum information entropy feature learning model - Google Patents
Rotary machine fault diagnosis method based on minimum information entropy feature learning model Download PDFInfo
- Publication number
- CN112861275A CN112861275A CN202110030109.XA CN202110030109A CN112861275A CN 112861275 A CN112861275 A CN 112861275A CN 202110030109 A CN202110030109 A CN 202110030109A CN 112861275 A CN112861275 A CN 112861275A
- Authority
- CN
- China
- Prior art keywords
- matrix
- fault
- vibration signal
- feature
- information entropy
- 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
- 238000000034 method Methods 0.000 title claims abstract description 39
- 238000003745 diagnosis Methods 0.000 title claims abstract description 23
- 239000011159 matrix material Substances 0.000 claims abstract description 91
- 238000012360 testing method Methods 0.000 claims abstract description 14
- 238000013507 mapping Methods 0.000 claims abstract description 7
- 238000005457 optimization Methods 0.000 claims abstract description 4
- 238000012549 training Methods 0.000 claims description 32
- 230000006870 function Effects 0.000 claims description 12
- 230000004913 activation Effects 0.000 claims description 6
- 238000010606 normalization Methods 0.000 claims description 5
- 238000000605 extraction Methods 0.000 abstract description 3
- 238000012544 monitoring process Methods 0.000 description 5
- 230000005540 biological transmission Effects 0.000 description 3
- 238000013135 deep learning Methods 0.000 description 2
- 238000011161 development Methods 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 238000001914 filtration Methods 0.000 description 2
- 238000013100 final test Methods 0.000 description 2
- 238000012423 maintenance Methods 0.000 description 2
- 238000005259 measurement Methods 0.000 description 2
- 230000000737 periodic effect Effects 0.000 description 2
- 238000005070 sampling Methods 0.000 description 2
- 238000002759 z-score normalization Methods 0.000 description 2
- 241001391944 Commicarpus scandens Species 0.000 description 1
- 238000005299 abrasion Methods 0.000 description 1
- 230000001133 acceleration Effects 0.000 description 1
- 230000010485 coping Effects 0.000 description 1
- 238000013480 data collection Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000001514 detection method Methods 0.000 description 1
- 230000007613 environmental effect Effects 0.000 description 1
- 239000000284 extract Substances 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 238000005096 rolling process Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/10—Geometric CAD
- G06F30/17—Mechanical parametric or variational design
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
- G06F18/2415—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on parametric or probabilistic models, e.g. based on likelihood ratio or false acceptance rate versus a false rejection rate
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
- G06F30/27—Design optimisation, verification or simulation using machine learning, e.g. artificial intelligence, neural networks, support vector machines [SVM] or training a model
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Evolutionary Computation (AREA)
- Geometry (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Artificial Intelligence (AREA)
- Data Mining & Analysis (AREA)
- Computer Hardware Design (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Medical Informatics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Software Systems (AREA)
- Evolutionary Biology (AREA)
- Probability & Statistics with Applications (AREA)
- Computational Mathematics (AREA)
- Mathematical Analysis (AREA)
- Mathematical Optimization (AREA)
- Pure & Applied Mathematics (AREA)
- Testing Of Devices, Machine Parts, Or Other Structures Thereof (AREA)
Abstract
The invention provides a rotary machine fault diagnosis method based on a minimum information entropy feature learning model, which comprises the following steps of: constructing a Hankel matrix, stacking the Hankel matrix, and obtaining a characteristic matrix through linear mapping; to rows of the feature matrixL 2Normalizing the norm, and then minimizing the information entropy of the feature matrix to obtain a trained weight matrix; solving the characteristics of the fault vibration signal through the trained weight matrix, standardizing the fault vibration signal by Z-score, combining the fault vibration signal with the label, and importing the fault vibration signal into a Softmax model for optimization to obtain a trained Softmax classifier; and (3) calculating the characteristics of the vibration signals of the rotary machine under the test environment through the weight matrix, standardizing the vibration signals by Z-score, and inputting the standardized vibration signals into a Softmax classifier to obtain a fault diagnosis result of the rotary machine. The invention is characterized in thatCompared with the existing unsupervised feature extraction method, the method has better noise adaptability, accuracy and robustness.
Description
Technical Field
The invention relates to the technical field of intelligent fault diagnosis of vibration signals, in particular to a rotary machine fault diagnosis method based on a minimum information entropy feature learning model.
Background
With the development of modern industry and the progress of science and technology, loading tools such as engineering vehicles, rail transit and the like are developing towards high speed, high precision and high efficiency, and establishing a reliable health monitoring system is a necessary measure for ensuring the high-efficiency and safe operation of the equipment, and has important significance for reasonably prolonging the service life of mechanical equipment, reducing the periodic maintenance cost and ensuring the safety of the operation of the equipment. As a main part for transmitting power, the gear box has the characteristics of compact structure, high transmission precision and large transmission torque, and is widely applied to the fields of aviation, engineering vehicles, rail transit and the like. The rotating parts in the gear box, such as a shaft, a bearing, a gear and the like, have the characteristics of high processing precision, small transmission error, high equipment precision and the like, work under various complex and changeable working conditions and running environments, are the most critical parts of the gear box and parts which are easy to break down, and fault signals are often submerged in environmental noise and are not easy to perceive.
The establishment of a reliable health monitoring system is a necessary measure for ensuring the high-efficiency and safe operation of the equipment, and has important significance for reasonably prolonging the service life of mechanical equipment, reducing periodic maintenance cost and ensuring the safety of the operation of the equipment. Due to the rapid development of computer networks, the scale of equipment groups for mechanical health monitoring is large, the number of required measuring points is large, the data collection duration is long, and massive data are acquired, so that the mechanical health monitoring system is promoted to enter a big data era. The traditional fault diagnosis method is incapable of coping with the fault, an advanced data driving method is researched and utilized, information is extracted from big mechanical data, the health condition of equipment is accurately and efficiently identified, and the fault diagnosis method becomes a new hotspot problem in the field of mechanical fault diagnosis. The mechanical health monitoring method based on deep learning realizes the organic combination of unsupervised learning and supervised learning, can simultaneously complete the self-adaptive extraction of big data fault characteristics and the identification of mechanical health conditions, and overcomes the defects of the traditional method in characteristic extraction and fault identification.
Most deep learning algorithms have a lot of hyper-parameters to be adjusted, the adjusting processes need to be experienced a priori, and meanwhile, the feature generalization capability in a noise environment is poor.
Disclosure of Invention
In order to solve the problems, the invention provides a rotary machine fault diagnosis method based on a minimum information entropy feature learning model.
In order to achieve the purpose, the technical scheme of the invention is as follows:
a rotary machine fault diagnosis method based on a minimum information entropy feature learning model comprises the following steps:
constructing a Hankel matrix according to the collected fault vibration signals, stacking the Hankel matrix, and then obtaining a characteristic matrix through linear mapping;
carry out L to the line of the characteristic matrix2Normalizing the norm, and then minimizing the information entropy of the feature matrix to obtain a trained weight matrix;
the method comprises the steps of solving the characteristics of fault vibration signals through a trained weight matrix, carrying out Z-score standardization on the characteristics of the fault vibration signals, combining the characteristics with labels, importing the characteristics into a Softmax model for optimization, and obtaining a Softmax classifier after training;
and calculating the characteristics of the rotary mechanical vibration signal under the test environment through a weight matrix, carrying out Z-score standardization on the characteristics of the rotary mechanical vibration signal, and inputting the characteristics into a Softmax classifier to obtain a fault diagnosis result of the rotary machine.
Further, the process of constructing the Hankel matrix in step 1 is as follows: setting M samples of the rotary mechanical vibration signal with the label asHankel matrix for constructing fault vibration signal according to rotating speed information of rotating machineryWherein N isinIs the length of the matrix array, NsIs the length of the matrix row; and stacking Hankel matrixes of all the rotating mechanical vibration samples and linearly mapping to obtain an input matrix S of the training model, wherein,
further, the process of minimizing the information entropy of the feature matrix in step 2 is: first minimizing the cost function J(W)So as to extract the distinctive features of the image,
wherein,for the normalized features of the ith row and the jth column in the input matrix, the row normalization mode of the feature matrix isfiThe characteristics of the ith row of the activated fault signal characteristic matrix; the feature activation function of the feature matrix is f ═ WS, where W is the weight matrix,Noutthe length of a matrix array in the weight matrix; and optimizing an objective function through an L-BFGS algorithm to train a weight matrix.
Further, the implementation process of training the Softmax classifier in step 3 is as follows: extracting the characteristics of the fault vibration signal through a characteristic activation function and a trained weight matrix; the features are then Z-score normalized, i.e. fz=(f-μf)-σfWherein f iszFor normalized features, f is the non-normalized fault vibration signal feature, μfFor non-normalized fault vibration signal characteristic fValue σfThe standard deviation of the non-normalized fault vibration signal characteristic f is obtained; adding a label to becomeWhereinNormalized feature of the ith sample, yiA failure label for the ith sample; and then training the Softmax training model to train a weight vector theta of the Softmax classifier so as to obtain the trained Softmax classifier.
The method extracts the sample characteristics through the minimum information entropy characteristic learning algorithm, and enables the fault diagnosis method to be more accurate and reasonable through the Z-score standardization of the Hankel matrix and the characteristics. The method comprises the steps of firstly training a minimum information entropy characteristic learning model through existing label data, then inputting a vibration signal obtained through testing into the trained model in the actual rotating machine health detection process, and carrying out fault diagnosis on the rotating machine. The method has better accuracy, robustness and noise adaptability.
Drawings
FIG. 1 is a flow chart of a method in an embodiment of the present invention.
FIG. 2 is a graph showing the comparison of the test effects in example 1 of the present invention.
FIG. 3 is a comparison chart of the test effects in example 2 of the present invention.
Detailed Description
The specific contents of the present invention will be further described with reference to the accompanying drawings and examples.
The embodiment of the invention provides a fault diagnosis method for a rotary machine, which comprises the following specific steps as shown in fig. 1:
constructing a Hankel matrix according to the collected fault vibration signals, stacking the Hankel matrix, and then obtaining a characteristic matrix through linear mapping;
carry out L to the line of the characteristic matrix2Normalizing the norm, and then minimizing the information entropy of the feature matrix to obtain a trained weight matrix;
the method comprises the steps of solving the characteristics of fault vibration signals through a trained weight matrix, carrying out Z-score standardization on the characteristics of the fault vibration signals, combining the characteristics with labels, importing the characteristics into a Softmax model for optimization, and obtaining a Softmax classifier after training;
and calculating the characteristics of the rotary mechanical vibration signal under the test environment through a weight matrix, carrying out Z-score standardization on the characteristics of the rotary mechanical vibration signal, and inputting the characteristics into a Softmax classifier to obtain a fault diagnosis result of the rotary machine.
The process of constructing the Hankel matrix in step 1 is as follows: setting M samples of the rotary mechanical vibration signal with the label asHankel matrix for constructing fault vibration signal according to rotating speed information of rotating machineryWherein N isinIs the length of the matrix array, NsIs the length of the matrix row; and stacking Hankel matrixes of all the rotating mechanical vibration samples and linearly mapping to obtain an input matrix S of the training model, wherein,
the process of minimizing the information entropy of the feature matrix in step 2 is: first minimizing the cost function J(W)So as to extract the distinctive features of the image,
wherein,for the normalized features of the ith row and the jth column in the input matrix, the row normalization mode of the feature matrix isfiThe characteristics of the ith row of the activated fault signal characteristic matrix; the feature activation function of the feature matrix is f ═ WS, where W is the weight matrix,Noutthe length of a matrix array in the weight matrix; and optimizing an objective function through an L-BFGS algorithm to train a weight matrix.
The implementation process of training the Softmax classifier in the step 3 is as follows: extracting the characteristics of the fault vibration signal through a characteristic activation function and a trained weight matrix; the features are then Z-score normalized, i.e. fz=(f-μf)-σfWherein f iszFor normalized features, f is the non-normalized fault vibration signal feature, μfIs the mean, σ, of the non-normalized fault vibration signal characteristic ffThe standard deviation of the non-normalized fault vibration signal characteristic f is obtained; adding a label to becomeWhereinNormalized feature of the ith sample, yiA failure label for the ith sample; and then training the Softmax training model to train a weight vector theta of the Softmax classifier so as to obtain the trained Softmax classifier.
Example 1:
and the data adopts bearing fault data disclosed by Kaiser university to carry out model training and testing. The fault types are { normal, inner ring fault (0.18mm, 0.36mm, 0.53mm), outer ring fault (0.18mm, 0.36mm, 0.53mm), rolling element fault (0.18mm, 0.36mm, 0.53mm) }, each fault type has four loads (1797r/min, 1772r/min, 1750r/min, 1730r/min), the sampling frequency is 12kHz data, the number of vibration sample points of each fault type is 1200, and the measurement is repeated 100 times. Thus a total of 4000 samples. 10% of these samples were taken as training samples, and the rest were test samples. First, set up trainingA sample matrix, wherein a Hankel matrix (the matrix length is 100) of each training sample is stacked to be used as a training sample matrix, training is carried out through a minimum entropy characteristic learning model, and the output dimension N isoutThe weight matrix W is trained to 100.
And multiplying the Hankel matrix of the training sample by W, carrying out Z-score normalization, combining with the label, carrying out Softmax model training, and testing by using the test sample after training. The accuracy of the final test reaches more than 99%. In order to illustrate the robustness of the intelligent fault diagnosis method based on the minimum information entropy feature learning model and the influence of the Hankel matrix and the feature normalization, the method, the minimum entropy model without feature normalization and the standard sparse filtering are compared, and the obtained result is shown in figure 2.
Example 2:
and the data adopts gearbox fault data to train and test the model. Data are collected by an acceleration sensor, and the sampling frequency is 12800 Hz. The fault types are { normal, abrasion, pitting and tooth breakage }, the vibration sample points of each fault type are 1280, and 100 times of repeated measurement are carried out. Thus there are a total of 400 samples. Firstly, establishing a training matrix sample matrix, stacking a Hankel matrix (the length of the matrix is 100) of each training sample as a training sample matrix, training through a minimum entropy characteristic learning model, and outputting a dimension NoutThe weight matrix W is trained to 100.
And multiplying the Hankel matrix of the training sample by W, carrying out Z-score normalization, combining with the label, carrying out the training of a Softmax feature classifier, and testing by using the test sample after the training is finished, wherein the accuracy of the final test reaches more than 99%.
In order to illustrate the accuracy and robustness of the proposed intelligent fault diagnosis method based on the minimum information entropy feature learning model, samples with different percentages are compared with standard sparse filtering as training sets, and the obtained result is shown in fig. 3.
Claims (4)
1. A rotary machine fault diagnosis method based on a minimum information entropy feature learning model is characterized by comprising the following steps:
constructing a Hankel matrix according to the collected fault vibration signals, stacking the Hankel matrix, and then obtaining a characteristic matrix through linear mapping;
carry out L to the line of the characteristic matrix2Normalizing the norm, and then minimizing the information entropy of the feature matrix to obtain a trained weight matrix;
the method comprises the steps of solving the characteristics of fault vibration signals through a trained weight matrix, carrying out Z-score standardization on the characteristics of the fault vibration signals, combining the characteristics with labels, importing the characteristics into a Softmax model for optimization, and obtaining a Softmax classifier after training;
and calculating the characteristics of the rotary mechanical vibration signal under the test environment through a weight matrix, carrying out Z-score standardization on the characteristics of the rotary mechanical vibration signal, and inputting the characteristics into a Softmax classifier to obtain a fault diagnosis result of the rotary machine.
2. The rotating machine fault diagnosis method based on the minimum information entropy feature learning model of claim 1, characterized in that: the process of constructing the Hankel matrix in step 1 is as follows: setting M samples of the rotary mechanical vibration signal with the label asHankel matrix for constructing fault vibration signal according to rotating speed information of rotating machineryWherein N isinIs a matrixLength of column, NsIs the length of the matrix row; and stacking Hankel matrixes of all the rotating mechanical vibration samples and linearly mapping to obtain an input matrix S of the training model, wherein,
3. the rotating machine fault diagnosis method based on the minimum information entropy feature learning model of claim 2, characterized in that: the process of minimizing the information entropy of the feature matrix in step 2 is: first minimizing the cost function J(W)So as to extract the distinctive features of the image,
wherein,for the normalized features of the ith row and the jth column in the input matrix, the row normalization mode of the feature matrix isfiThe characteristics of the ith row of the activated fault signal characteristic matrix; the feature activation function of the feature matrix is f ═ WS, where W is the weight matrix,Noutthe length of a matrix array in the weight matrix; and optimizing an objective function through an L-BFGS algorithm to train a weight matrix.
4. The rotating machine fault diagnosis method based on the minimum information entropy feature learning model of claim 3, characterized in that: the implementation process of training the Softmax classifier in the step 3 is as follows: extracting the characteristics of the fault vibration signal through the characteristic activation function and the trained weight matrix(ii) a The features are then Z-score normalized, i.e. fz=(f-μf)/σfWherein f iszF is the features of the fault vibration signal which is not normalized, carry forwardfIs the mean, σ, of the non-normalized fault vibration signal characteristic ffThe standard deviation of the non-normalized fault vibration signal characteristic f is obtained; adding a label to becomeWhereinNormalized feature of the ith sample, yiA failure label for the ith sample; and then training the Softmax training model to train a weight vector theta of the Softmax classifier so as to obtain the trained Softmax classifier.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202110030109.XA CN112861275A (en) | 2021-01-11 | 2021-01-11 | Rotary machine fault diagnosis method based on minimum information entropy feature learning model |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202110030109.XA CN112861275A (en) | 2021-01-11 | 2021-01-11 | Rotary machine fault diagnosis method based on minimum information entropy feature learning model |
Publications (1)
Publication Number | Publication Date |
---|---|
CN112861275A true CN112861275A (en) | 2021-05-28 |
Family
ID=76002270
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202110030109.XA Pending CN112861275A (en) | 2021-01-11 | 2021-01-11 | Rotary machine fault diagnosis method based on minimum information entropy feature learning model |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN112861275A (en) |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN113449783A (en) * | 2021-06-17 | 2021-09-28 | 广州大学 | Countermeasure sample generation method, system, computer device and storage medium |
CN113743585A (en) * | 2021-08-17 | 2021-12-03 | 山东科技大学 | Rotary machine early fault diagnosis method based on rapid nonlinear blind deconvolution |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
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 |
CN108225750A (en) * | 2017-12-12 | 2018-06-29 | 南京航空航天大学 | A kind of rotary machinery fault diagnosis method based on the more correlations of fault signature |
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 |
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 |
-
2021
- 2021-01-11 CN CN202110030109.XA patent/CN112861275A/en active Pending
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
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 |
CN108225750A (en) * | 2017-12-12 | 2018-06-29 | 南京航空航天大学 | A kind of rotary machinery fault diagnosis method based on the more correlations of fault signature |
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 |
Cited By (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN113449783A (en) * | 2021-06-17 | 2021-09-28 | 广州大学 | Countermeasure sample generation method, system, computer device and storage medium |
CN113743585A (en) * | 2021-08-17 | 2021-12-03 | 山东科技大学 | Rotary machine early fault diagnosis method based on rapid nonlinear blind deconvolution |
CN113743585B (en) * | 2021-08-17 | 2023-10-13 | 山东科技大学 | Rotary machine early fault diagnosis method based on fast nonlinear blind deconvolution |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN110110768B (en) | Rolling bearing fault diagnosis method based on parallel feature learning and multiple classifiers | |
CN111721535A (en) | Bearing fault detection method based on convolution multi-head self-attention mechanism | |
CN110334764B (en) | Rotary machine intelligent fault diagnosis method based on integrated depth self-encoder | |
CN113469060A (en) | Multi-sensor fusion convolution neural network aeroengine bearing fault diagnosis method | |
CN103034170B (en) | Numerical control machine tool machining performance prediction method based on intervals | |
CN108871762A (en) | A kind of gearbox of wind turbine method for diagnosing faults | |
CN111562108A (en) | Rolling bearing intelligent fault diagnosis method based on CNN and FCMC | |
CN112861275A (en) | Rotary machine fault diagnosis method based on minimum information entropy feature learning model | |
CN113008559B (en) | Bearing fault diagnosis method and system based on sparse self-encoder and Softmax | |
CN116793682A (en) | Bearing fault diagnosis method based on iCORAL-MMD and anti-migration learning | |
CN113188794B (en) | Gearbox fault diagnosis method and device based on improved PSO-BP neural network | |
CN112257862B (en) | Semi-supervised identification method based on relation network marking sample expansion | |
CN113505664B (en) | Fault diagnosis method for planetary gear box of wind turbine generator | |
CN113128338A (en) | Intelligent diagnosis method for printing machine roller fault under small sample | |
CN114626625B (en) | Bearing life prediction method based on multi-module U-BiLSTM network | |
CN113869339A (en) | Deep learning classification model for fault diagnosis and fault diagnosis method | |
CN112949402A (en) | Fault diagnosis method for planetary gear box under minimum fault sample size | |
CN114429152A (en) | Rolling bearing fault diagnosis method based on dynamic index antagonism self-adaption | |
CN113790890B (en) | Wavelet packet decomposition weight fuzzy entropy and ELM-based bearing fault classification method and device | |
Yang et al. | Self-attention parallel fusion network for wind turbine gearboxes fault diagnosis | |
Xu et al. | Fault diagnosis of subway traction motor bearing based on information fusion under variable working conditions | |
CN110160781B (en) | Test set reconstruction and prediction method for rotary machine fault classification | |
CN112163630A (en) | Compound fault diagnosis method and device based on unbalanced learning | |
CN113935403B (en) | Multi-label multi-classification based spindle bearing and cutter compound fault diagnosis method | |
CN111523557A (en) | Wind power intelligent fault diagnosis method based on big data |
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 |