CN113065418A - Rolling bearing fault diagnosis method based on SSA-WDCNN - Google Patents

Rolling bearing fault diagnosis method based on SSA-WDCNN Download PDF

Info

Publication number
CN113065418A
CN113065418A CN202110291325.XA CN202110291325A CN113065418A CN 113065418 A CN113065418 A CN 113065418A CN 202110291325 A CN202110291325 A CN 202110291325A CN 113065418 A CN113065418 A CN 113065418A
Authority
CN
China
Prior art keywords
rolling bearing
matrix
fault diagnosis
wdcnn
ssa
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
Application number
CN202110291325.XA
Other languages
Chinese (zh)
Inventor
朱瑞
王明鑫
徐思宇
韩清鹏
夏鑫
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Shanghai University of Electric Power
Shanghai Electric Power University
University of Shanghai for Science and Technology
Original Assignee
Shanghai Electric Power University
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Shanghai Electric Power University filed Critical Shanghai Electric Power University
Priority to CN202110291325.XA priority Critical patent/CN113065418A/en
Publication of CN113065418A publication Critical patent/CN113065418A/en
Pending legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
    • G06F2218/12—Classification; Matching
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01M—TESTING STATIC OR DYNAMIC BALANCE OF MACHINES OR STRUCTURES; TESTING OF STRUCTURES OR APPARATUS, NOT OTHERWISE PROVIDED FOR
    • G01M13/00—Testing of machine parts
    • G01M13/04—Bearings
    • G01M13/045—Acoustic or vibration analysis
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/08—Learning methods
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
    • G06F2218/02—Preprocessing
    • G06F2218/04—Denoising
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
    • G06F2218/08—Feature extraction

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Artificial Intelligence (AREA)
  • General Engineering & Computer Science (AREA)
  • Biomedical Technology (AREA)
  • General Health & Medical Sciences (AREA)
  • Health & Medical Sciences (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Software Systems (AREA)
  • Biophysics (AREA)
  • Computational Linguistics (AREA)
  • Data Mining & Analysis (AREA)
  • Evolutionary Computation (AREA)
  • Mathematical Physics (AREA)
  • Molecular Biology (AREA)
  • Computing Systems (AREA)
  • Signal Processing (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Acoustics & Sound (AREA)
  • Testing Of Devices, Machine Parts, Or Other Structures Thereof (AREA)
  • Measurement Of Mechanical Vibrations Or Ultrasonic Waves (AREA)

Abstract

The invention relates to a rolling bearing fault diagnosis method based on SSA-WDCNN, which comprises the following steps: s1: performing singular spectrum analysis on the rolling bearing vibration signal containing noise to obtain a reconstructed vibration signal; s2: inputting the reconstructed vibration signal into a fault diagnosis model for training; s3: and sending the original signal to be diagnosed into a trained fault diagnosis model to diagnose the fault of the rolling bearing. Compared with the prior art, the rolling bearing vibration signal containing noise is processed by using a singular spectrum analysis method, and the bearing vibration signal is diagnosed based on the convolutional neural network with the first convolutional layer as the wide convolutional layer, so that the extraction effect and accuracy of signal characteristics are improved, and the diagnosis accuracy and diagnosis efficiency of the rolling bearing fault are further improved.

Description

Rolling bearing fault diagnosis method based on SSA-WDCNN
Technical Field
The invention relates to the field of fault diagnosis, in particular to a rolling bearing fault diagnosis method based on SSA-WDCNN.
Background
The rolling bearing is one of key parts in the rotating machinery, the health condition of the rolling bearing influences the normal operation of the whole rotating machinery system, and the loss in production can be effectively reduced by timely finding the fault of the bearing. At present, vibration signals of rolling bearings are mostly researched aiming at fault diagnosis of the rolling bearings, however, fault characteristic components in the signals, particularly early weak fault characteristics, are easily submerged by noise and other irrelevant signal components, so that equipment faults cannot be discovered in time. Therefore, the signal decomposition method is used for decomposing the vibration signal into a series of sub-signal components with clear physical significance, and further extracting fault characteristic components from the sub-signal components, which is particularly important for mechanical fault diagnosis.
In the prior art, a self-encoder is used for carrying out noise reduction on a noisy signal and then inputting the signal into a One-dimensional convolutional neural network (One-Dimension CNN,1-DCNN) model to realize feature extraction of a fault signal, but training of two models is required and only a test signal with a signal-to-noise ratio smaller than-1 dB is used. In addition, partial interference of the original signal is eliminated by utilizing wavelet packet decomposition, and then Empirical Mode Decomposition (EMD) is adopted to extract fault features, so that double-spectrum demodulation of the whole signal is realized, but the resolution of the final double-spectrum is not high. In addition, the method also has application in signal noise reduction for integrated empirical mode decomposition (EEMD) after EMD improvement, and in the prior art, a relatively good test result is obtained by denoising a vibration signal by combining an EEMD method and a block threshold strategy, but a reconstructed signal of the method has a large deviation from a clean signal, and the application range is limited to a small signal-to-noise ratio signal. In the existing related researches, the effects of noise reduction and signal feature extraction on large noise are not ideal all the time, so that some fault signals are very easy to generate diagnosis deviation under the interference of a large noise environment.
Disclosure of Invention
The invention aims to overcome the defects of the prior art and provide a rolling bearing fault diagnosis method based on SSA-WDCNN.
The purpose of the invention can be realized by the following technical scheme:
a rolling bearing fault diagnosis method based on SSA-WDCNN comprises the following steps:
s1: performing singular spectrum analysis on the rolling bearing vibration signal containing noise to obtain a reconstructed vibration signal;
s2: inputting the reconstructed vibration signal into a fault diagnosis model for training;
s3: and sending the original signal to be diagnosed into a trained fault diagnosis model to diagnose the fault of the rolling bearing.
SSA, Singular spectral Analysis, WDCNN, is a finger-wide convolution kernel depth neural network.
Preferably, the fault diagnosis model is a convolutional neural network.
Preferably, the first convolutional layer of the convolutional neural network is a wide convolutional layer.
Preferably, the convolutional neural network includes a first convolutional layer, a BN layer, a first pooling layer, a second convolutional layer, a second pooling layer, a third convolutional layer, a third pooling layer, a fourth convolutional layer, a fourth pooling layer, a fifth convolutional layer, a fifth pooling layer, a full-link layer, and a softmax layer, which are sequentially arranged, where the first convolutional layer is a wide convolutional layer.
Preferably, the wide convolutional layer has a convolution kernel width of 64.
Preferably, step S1 specifically includes:
s11: converting a vibration signal of the rolling bearing containing noise into a track matrix;
s12: performing singular value decomposition on the track matrix;
s13: reconstructing a new track matrix according to the singular value decomposition result;
s14: carrying out diagonalization average on the new track matrix obtained by reconstruction, and converting the new track matrix into a one-dimensional time sequence;
s15: and judging whether the signal-to-noise ratio of the one-dimensional time sequence of the new track matrix is greater than a set signal-to-noise ratio, if so, entering a step S2, and otherwise, sending the one-dimensional time sequence to the step S11.
Preferably, the rolling bearing vibration signal X containing noise in step S11NComprises the following steps:
XN＝{f1,f2，…，fN}
wherein f isNFor time series signal components of length N,
the transformed trajectory matrix X is:
Figure BDA0002982802220000031
wherein f isn+m-1Is the element signal of the m × n window trace matrix.
Preferably, the specific step of step S12 includes:
defining covariance matrix C-XXTAnd solving the eigenvalues of the matrix C and arranging the eigenvalues in a descending order according to the magnitude of the eigenvalues to obtain m eigenvalues of the matrix C: lambda [ alpha ]1≥λ2≥...≥λmMore than or equal to 0, the eigenvector corresponding to the m eigenvalues is U1，U2…Um，
Definition of
Figure BDA0002982802220000032
The singular value decomposition of the trajectory matrix X is as follows:
X＝X1+X2+…Xi+…Xm
wherein the content of the first and second substances,
Figure BDA0002982802220000033
Figure BDA0002982802220000034
being singular values of the matrix X, UiAs a verified cross function of the trajectory matrix, ViAs a major component, ViAs the right eigenvector of the trajectory matrix, UiIs the left eigenvector of the trajectory matrix.
Preferably, the specific step of step S13 includes:
setting the order d of the singular value to be extracted, determining the value d by adopting a singular value center difference quotient method when d is less than m, and determining the center difference quotient Z when the singular value is the d-th valuedObtaining maximum value, obtaining the component of singular value decomposition from first order to d order, setting other singular value components as zero, selecting multiple orders from the extracted d order singular value components for reconstruction to obtain new singular value matrix, combining with
Figure BDA0002982802220000035
A new trajectory matrix may be obtained.
Preferably, the central difference quotient is ZiComprises the following steps:
Figure BDA0002982802220000036
the new trajectory matrix Y is:
Y＝X1+...+Xi。
preferably, in step S14, the new trajectory matrix Y is converted into a one-dimensional time sequence Yk：
Figure BDA0002982802220000037
Wherein k is the new one-dimensional time series element order, and p is the number of diagonal elements averaged by the new trajectory matrix.
Compared with the prior art, the invention has the following advantages:
(1) the rolling bearing vibration signal containing noise is processed by using a singular spectrum analysis method, and the bearing vibration signal is diagnosed based on the convolutional neural network with the first convolutional layer as the wide convolutional layer, so that fault characteristic components in the signal, particularly early weak fault characteristics, are prevented from being easily submerged by noise and other irrelevant signal components, the extraction effect and accuracy of the signal characteristics are improved, and the diagnosis accuracy and diagnosis efficiency of the rolling bearing fault are further improved;
(2) the method utilizes a singular spectrum analysis method to perform noise reduction on the vibration signal, widens the range of basic noise reduction, ensures the practical application capability of a diagnosis model to large noise and multiple working conditions, and has good signal noise reduction effect;
(3) the invention diagnoses the bearing vibration signal by adopting the convolutional neural network with the first convolutional layer as the wide convolutional layer, extracts the short-term characteristics by using the large-size convolutional kernel, extracts the effective fault characteristics of the original signal to the maximum extent, abandons the other characteristics which are relatively ineffective, and expands the whole network depth by selecting the small convolutional kernel for the subsequent convolutional layer, thereby preventing overfitting, improving the characteristic acquisition effect of the convolutional neural network and improving the diagnosis accuracy of the bearing fault.
Drawings
FIG. 1 is a flow chart of the present invention;
FIG. 2 is a block diagram of a convolutional neural network of the present invention;
FIG. 3 is a graph of an error matrix for fault diagnosis prediction in accordance with the present invention;
FIG. 4 shows the last layer of convolutional layer feature extraction during the model training process of the present invention.
Detailed Description
The invention is described in detail below with reference to the figures and specific embodiments. Note that the following description of the embodiments is merely a substantial example, and the present invention is not intended to be limited to the application or the use thereof, and is not limited to the following embodiments.
Examples
A rolling bearing fault diagnosis method based on SSA-WDCNN comprises the following steps:
s1: and carrying out singular spectrum analysis on the rolling bearing vibration signal containing the noise to obtain a reconstructed vibration signal.
Step S1 specifically includes:
s11: and converting the vibration signal of the rolling bearing containing the noise into a track matrix.
Rolling bearing vibration signal X containing noise in step S11NComprises the following steps:
XN＝{f1,f2，…，fN}
wherein f isNFor time series signal components of length N,
the transformed trajectory matrix X is:
Figure BDA0002982802220000051
wherein f isn+m-1Is the element signal of the m × n window trace matrix.
S12: and carrying out singular value decomposition on the track matrix.
The specific steps of step S12 include:
defining covariance matrix C-XXTAnd solving the eigenvalues of the matrix C and arranging the eigenvalues in a descending order according to the magnitude of the eigenvalues to obtain m eigenvalues of the matrix C: lambda [ alpha ]1≥λ2≥...≥λmMore than or equal to 0, the eigenvector corresponding to the m eigenvalues is U1，U2…Um，
Definition of
Figure BDA0002982802220000052
The singular value decomposition of the trajectory matrix X is as follows:
X＝X1+X2+…Xi+…Xm
wherein the content of the first and second substances,
Figure BDA0002982802220000053
Figure BDA0002982802220000054
being singular values of the matrix X, UiAs a verified cross function of the trajectory matrix, ViAs a major component, ViAs the right eigenvector of the trajectory matrix, UiIs the left eigenvector of the trajectory matrix.
S13: and reconstructing a new track matrix according to the singular value decomposition result.
The specific steps of step S13 include:
for order of singular value to be extractedd is set, d is less than m, d value is confirmed by singular value center difference quotient method, when the singular value is the d-th one, the center difference quotient Z isdObtaining maximum value, obtaining the component of singular value decomposition from first order to d order, setting other singular value components as zero, selecting multiple orders from the extracted d order singular value components for reconstruction to obtain new singular value matrix, combining with
Figure BDA0002982802220000055
A new trajectory matrix may be obtained.
The center difference quotient is ZiComprises the following steps:
Figure BDA0002982802220000056
the new trajectory matrix Y is:
Y＝X1+...+Xi。
s14: and carrying out diagonalization and averaging on the track matrix obtained by reconstruction, and converting the new track matrix into a one-dimensional time sequence.
In step S14, the new trajectory matrix Y is converted into a one-dimensional time series Yk：
Figure BDA0002982802220000061
Wherein k is the new one-dimensional time series element order, and p is the number of diagonal elements averaged by the new trajectory matrix.
S15: and judging whether the signal-to-noise ratio of the one-dimensional time sequence of the new track matrix is greater than a set signal-to-noise ratio, if so, entering a step S2, and otherwise, sending the one-dimensional time sequence to the step S11. In this embodiment, the set snr is-1 dB.
S2: and inputting the reconstructed vibration signal into a fault diagnosis model for training.
In the invention, the fault diagnosis model is a convolutional neural network.
Specifically, the fault diagnosis model in the invention is WDCNN, i.e. a wide convolutional kernel deep neural network, the first convolutional layer of which is a wide convolutional layer.
In this embodiment, as shown in fig. 2, the reconstructed vibration signal is directly input to a convolutional neural network model with a wide convolutional kernel to perform feature extraction and training, and the whole model weight is updated and optimized through error back propagation. The first layer of convolutional layer adopts large-size convolutional kernels to extract short-time features, effective fault feature extraction is carried out on original signals to the maximum extent, other relatively ineffective features are abandoned, and the subsequent convolutional layers adopt small convolutional kernels to expand the depth of the whole network and prevent overfitting. Specifically, the convolutional neural network comprises a first convolutional layer, a BN layer, a first pooling layer, a second convolutional layer, a second pooling layer, a third convolutional layer, a third pooling layer, a fourth convolutional layer, a fourth pooling layer, a fifth convolutional layer, a fifth pooling layer, a full-link layer and a softmax layer which are arranged in sequence, wherein the first convolutional layer is a wide convolutional layer, and the core width of the first convolutional layer is 64.
Specifically, as shown in table 1 below, the structural parameters of the whole convolutional neural network of the present invention are shown.
TABLE 1 convolutional neural network architecture parameter Table
Figure BDA0002982802220000071
S3: and sending the original signal to be diagnosed into a trained fault diagnosis model to diagnose the fault of the rolling bearing.
Specifically, in this embodiment, the trained model is used to perform specific fault diagnosis on the original signal, including the diagnosis accuracy test of the overall fault in the multi-operating-condition and high-noise environment, where the error of the diagnosis test on the fault signal with the operating-condition rotating speed of 1797rpm and-10 dB is shown in fig. 3, and the diagnosis accuracy of various operating conditions is shown in table 2, it can be seen that when the fault diagnosis is performed by using the method of the present invention, the overall diagnosis accuracy is over 93%, and the denoising effect and the diagnosis accuracy of the model can be ensured for large noise, and the overall method has good multi-operating-condition adaptability and practicability.
TABLE 2 model Fault diagnosis accuracy under multiple operating conditions
Figure BDA0002982802220000081
The above embodiments are merely examples and do not limit the scope of the present invention. These embodiments may be implemented in other various manners, and various omissions, substitutions, and changes may be made without departing from the technical spirit of the present invention.

Claims (10)

1. A rolling bearing fault diagnosis method based on SSA-WDCNN is characterized by comprising the following steps:
s1: performing singular spectrum analysis on the rolling bearing vibration signal containing noise to obtain a reconstructed vibration signal;
s2: inputting the reconstructed vibration signal into a fault diagnosis model for training;
s3: and sending the original signal to be diagnosed into a trained fault diagnosis model to diagnose the fault of the rolling bearing.
2. The SSA-WDCNN based rolling bearing fault diagnosis method of claim 1, wherein the fault diagnosis model is a convolutional neural network.
3. The SSA-WDCNN-based rolling bearing fault diagnosis method of claim 2, wherein the first convolutional layer of the convolutional neural network is a wide convolutional layer.
4. The SSA-WDCNN-based rolling bearing fault diagnosis method of claim 2, wherein the convolutional neural network comprises a first convolutional layer, a BN layer, a first pooling layer, a second convolutional layer, a second pooling layer, a third convolutional layer, a third pooling layer, a fourth convolutional layer, a fourth pooling layer, a fifth convolutional layer, a fifth pooling layer, a fully-connected layer and a softmax layer which are sequentially arranged, wherein the first convolutional layer is a wide convolutional layer.
5. The SSA-WDCNN-based rolling bearing fault diagnosis method of claim 4, wherein the wide convolution layer has a convolution kernel width of 64.
6. The SSA-WDCNN-based rolling bearing fault diagnosis method of claim 1, wherein the step S1 specifically comprises:
s11: converting a vibration signal of the rolling bearing containing noise into a track matrix;
s12: performing singular value decomposition on the track matrix;
s13: reconstructing a new track matrix according to the singular value decomposition result;
s14: carrying out diagonalization average on the new track matrix obtained by reconstruction, and converting the new track matrix into a one-dimensional time sequence;
s15: and judging whether the signal-to-noise ratio of the one-dimensional time sequence of the new track matrix is greater than a set signal-to-noise ratio, if so, entering a step S2, and otherwise, sending the one-dimensional time sequence to the step S11.
7. The SSA-WDCNN-based rolling bearing fault diagnosis method as claimed in claim 6, wherein the step S11 includes a vibration signal X of the rolling bearing containing noiseNComprises the following steps:
XN＝{f1,f2，…，fN}
wherein f isNFor time series signal components of length N,
the transformed trajectory matrix X is:
Figure FDA0002982802210000021
wherein f isn+m-1Is the element signal of the m × n window trace matrix.
8. The SSA-WDCNN based rolling bearing fault diagnosis method of claim 7, wherein the specific steps of step S12 include:
defining covariance matrix C-XXTAnd solving the eigenvalues of the matrix C and arranging the eigenvalues in a descending order according to the magnitude of the eigenvalues to obtain m eigenvalues of the matrix C: lambda [ alpha ]1≥λ2≥...≥λmMore than or equal to 0, the eigenvector corresponding to the m eigenvalues is U1，U2…Um，
Definition of
Figure FDA0002982802210000022
The singular value decomposition of the trajectory matrix X is as follows:
X＝X1+X2+…Xi+…Xm
wherein the content of the first and second substances,
Figure FDA0002982802210000023
Figure FDA0002982802210000024
being singular values of the matrix X, UiAs a verified cross function of the trajectory matrix, ViAs a major component, ViAs the right eigenvector of the trajectory matrix, UiIs the left eigenvector of the trajectory matrix.
9. The SSA-WDCNN based rolling bearing fault diagnosis method of claim 8, wherein the specific steps of step S13 include:
setting the order d of the singular value to be extracted, determining the value d by adopting a singular value center difference quotient method when d is less than m, and determining the center difference quotient Z when the singular value is the d-th valuedObtaining maximum value, obtaining the component of singular value decomposition from first order to d order, setting other singular value components as zero, selecting multiple orders from the extracted d order singular value components for reconstruction to obtain new singular value matrix, combining with
Figure FDA0002982802210000025
A new trajectory matrix may be obtained.
10. The SSA-WDCNN-based rolling bearing fault diagnosis method as claimed in claim 1, wherein the new trajectory matrix Y is converted into a one-dimensional time series Y in step S14k：
Figure FDA0002982802210000026
Wherein k is the new one-dimensional time series element order, and p is the number of diagonal elements averaged by the new trajectory matrix.
CN202110291325.XA 2021-03-18 2021-03-18 Rolling bearing fault diagnosis method based on SSA-WDCNN Pending CN113065418A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202110291325.XA CN113065418A (en) 2021-03-18 2021-03-18 Rolling bearing fault diagnosis method based on SSA-WDCNN

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202110291325.XA CN113065418A (en) 2021-03-18 2021-03-18 Rolling bearing fault diagnosis method based on SSA-WDCNN

Publications (1)

Publication Number Publication Date
CN113065418A true CN113065418A (en) 2021-07-02

Family

ID=76562027

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202110291325.XA Pending CN113065418A (en) 2021-03-18 2021-03-18 Rolling bearing fault diagnosis method based on SSA-WDCNN

Country Status (1)

Country Link
CN (1) CN113065418A (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN116662793A (en) * 2023-07-31 2023-08-29 昆仑数智科技有限责任公司 Rotor crack fault diagnosis and model construction method and device

Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN106226407A (en) * 2016-07-25 2016-12-14 中国电子科技集团公司第二十八研究所 A kind of online preprocess method of ultrasound echo signal based on singular spectrum analysis
CN106650071A (en) * 2016-12-12 2017-05-10 中国航空工业集团公司上海航空测控技术研究所 Intelligent fault diagnosis method for rolling bearing
CN107121285A (en) * 2017-04-21 2017-09-01 南京理工大学 A kind of bearing vibration signal fault feature extracting method
CN108489719A (en) * 2018-04-09 2018-09-04 常州湖南大学机械装备研究院 A kind of combined failure of rotating machinery diagnostic method based on the unusual spectral factorizations of G-P
CN109583386A (en) * 2018-11-30 2019-04-05 中南大学 A kind of intelligent rotating mechanical breakdown depth network characterization discrimination method
CN111222458A (en) * 2020-01-06 2020-06-02 浙江工业大学 Rolling bearing fault diagnosis method based on ensemble empirical mode decomposition and convolutional neural network

Patent Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN106226407A (en) * 2016-07-25 2016-12-14 中国电子科技集团公司第二十八研究所 A kind of online preprocess method of ultrasound echo signal based on singular spectrum analysis
CN106650071A (en) * 2016-12-12 2017-05-10 中国航空工业集团公司上海航空测控技术研究所 Intelligent fault diagnosis method for rolling bearing
CN107121285A (en) * 2017-04-21 2017-09-01 南京理工大学 A kind of bearing vibration signal fault feature extracting method
CN108489719A (en) * 2018-04-09 2018-09-04 常州湖南大学机械装备研究院 A kind of combined failure of rotating machinery diagnostic method based on the unusual spectral factorizations of G-P
CN109583386A (en) * 2018-11-30 2019-04-05 中南大学 A kind of intelligent rotating mechanical breakdown depth network characterization discrimination method
CN111222458A (en) * 2020-01-06 2020-06-02 浙江工业大学 Rolling bearing fault diagnosis method based on ensemble empirical mode decomposition and convolutional neural network

Non-Patent Citations (2)

* Cited by examiner, † Cited by third party
Title
董绍江等: "基于多层降噪技术及改进卷积神经网络的滚动轴承故障诊断方法", 《机械工程学报》 *
陈全: "深度学习及其在地电场异常检测中的应用研究", 《中国优秀博硕士学位论文全文数据库(硕士)基础科学辑》 *

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN116662793A (en) * 2023-07-31 2023-08-29 昆仑数智科技有限责任公司 Rotor crack fault diagnosis and model construction method and device
CN116662793B (en) * 2023-07-31 2024-03-08 昆仑数智科技有限责任公司 Rotor crack fault diagnosis and model construction method and device

Similar Documents

Publication Publication Date Title
CN111089726B (en) Rolling bearing fault diagnosis method based on optimal dimension singular spectrum decomposition
CN110672343B (en) Rotary machine fault diagnosis method based on multi-attention convolutional neural network
CN111238814B (en) Rolling bearing fault diagnosis method based on short-time Hilbert transform
CN111814656B (en) Electrocardiosignal noise reduction method based on countermeasure generation network
CN109211546B (en) Rotary machine fault diagnosis method based on noise reduction automatic encoder and increment learning
CN111160167A (en) Spindle fault classification and identification method based on S-transform deep convolutional neural network
Feng et al. Normalized wavelet packets quantifiers for condition monitoring
CN107817106A (en) Fault Diagnosis of Roller Bearings based on Bayes's residual transform singular value decomposition Gaussian Mixture HMM framework
CN113375940A (en) Fault bearing diagnosis method based on SVD and CEEMDAN
CN112945557B (en) Slewing bearing fault diagnosis method, slewing bearing fault diagnosis device and storage medium
CN113435322B (en) Method, system, equipment and readable storage medium for detecting faults of spindle bearing
CN112183259B (en) Rolling bearing fault diagnosis method based on CEEMD and kurtosis weighted average threshold denoising
CN113375939A (en) Mechanical part fault diagnosis method based on SVD and VMD
CN106502850B (en) A kind of the multiple faults discrimination method and device of sparse tight frame dictionary learning model
CN114112400A (en) Mechanical bearing fault diagnosis method based on multi-angle information fusion
CN111076934A (en) Method for diagnosing potential fault of bearing based on S transformation
CN113065418A (en) Rolling bearing fault diagnosis method based on SSA-WDCNN
CN113776837B (en) Rolling bearing fault diagnosis method based on CEEMDAN and GWO-NLM
CN113865872B (en) Bearing fault diagnosis method based on wavelet packet reconstruction imaging and CNN
CN114964782A (en) Rolling bearing fault diagnosis acoustic emission signal processing method and system based on compressed sensing
CN115062665A (en) Rolling bearing early fault diagnosis method based on self-adaptive variational modal decomposition
CN115452377B (en) Rolling bearing fault diagnosis method based on CEEMDAN and adaptive denoising
CN113281047A (en) Bearing inner and outer ring fault quantitative trend diagnosis method based on variable-scale Lempel-Ziv
Yu et al. Sparse time–frequency representation for the transient signal based on low-rank and sparse decomposition
CN114964783B (en) Gearbox fault detection model based on VMD-SSA-LSSVM

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